Table Of Contents
Caveats for the Cisco uBR10012 Universal Broadband Router in Cisco IOS Release 12.2SC
Open Caveats—Cisco IOS Release 12.2(33)SCB4
Resolved Caveats — Cisco IOS Release 12.2(33)SCB4
Open Caveats—Cisco IOS Release 12.2(33)SCB3
Resolved Caveats — Cisco IOS Release 12.2(33)SCB3
Open Caveats—Cisco IOS Release 12.2(33)SCB2
Resolved Caveats — Cisco IOS Release 12.2(33)SCB2
Open Caveats — Cisco IOS Release 12.2(33)SCB1
Resolved Caveats — Cisco IOS Release 12.2(33)SCB1
Open Caveats — Cisco IOS Release 12.2(33)SCB
Resolved Caveats — Cisco IOS Release 12.2(33)SCB
Open Caveats — Cisco IOS Release 12.2(33)SCA2
Resolved Caveats — Cisco IOS Release 12.2(33)SCA2
Open Caveats — Cisco IOS Release 12.2(33)SCA1
Resolved Caveats — Cisco IOS Release 12.2(33)SCA1
Open Caveats — Cisco IOS Release 12.2(33)SCA
Resolved Caveats — Cisco IOS Release 12.2(33)SCA
Obtaining Documentation and Submitting a Service Request
Caveats for the Cisco uBR10012 Universal Broadband Router in Cisco IOS Release 12.2SC
September 23, 2009
OL-15874-08
This document lists severity 1 and 2 caveats and select severity 3 caveats for the Cisco uBR10012 Universal Broadband Router in Cisco IOS Release 12.2SC, up to and including Cisco IOS Release 12.2(33)SCB4. Caveats describe unexpected behavior or defects in Cisco IOS software releases. Severity 1 caveats are the most serious caveats; severity 2 caveats are less serious. Severity 3 caveats are moderate caveats, and only selected severity 3 caveats are included in the caveats document.
Use this caveats document with the following release notes available on Cisco.com:
•
Cross-Platform Release Notes for Cisco Universal Broadband Routers in Cisco IOS Release 12.2SC
•
•
Note
Caveats
This document describes open and resolved severity 1 and 2 caveats and select severity 3 caveats:
•
•
Within the sections, the caveats are sorted by component in alphabetical order. The caveats are also sorted alphanumerically by caveat number.
The following information is provided for each caveat:
•
•
•
Open Caveats—Cisco IOS Release 12.2(33)SCB4
•
Symptoms: Wideband modem downstream throughput rate is not yet supported.
Workaround: There is no workaround.
•
Symptoms: The Cisco uBR10012 router may generate the following message:
UBR10000-6-BADIPASSIGNMENT: DHCPOFFER dropped : Interface Cablex/y/zMac abcd.efgh.ijkl SID 2205 L3_Interface Cablex/y/z IP x.y.z.talready assigned to MAC=mnop.qrst.uvwx Interface Cablex/y/z SID 6367Conditions: The cable modem may be stuck in init(d) for some hours before coming online. The Cable Modems are provisioned by an external DHCP server.
Workaround: If it is not possible to wait for the CM to come up, then a cable card reset will fix the issue.
•
Symptoms: In syntax check mode, if there is a standby in SSO mode, the cts dot1x command does not work.
Workaround: There are no workarounds.
•
Symptoms: The throughput rate shows (slightly) more than the expected range.
Conditions: This happens when the the DMIC feature is enabled and the modem is marked with given maximum rate values. This can also happen when the QoS profile has lower maximum rates.
Workaround: There are no workarounds.
•
Symptoms: 12.3(23)BC2 mixed docsis mode (atdma-tdma) can only achieve 40 UGS for an upstream. But, 12.3(17b)BC9 can achieve 48 UGS flows.
Conditions: Currently, it looks like this may have something to do with mixed mode and modems.
Workaround: There is no workaround.
•
Symptoms: Multicast stream drop, after changing the "ip igmp query-interval". The following show the output before and after the change:
Before the change:
Router# show ip mroute<<..skipped..>>(, ), 01:09:04/00:02:52, flags: sTIIncoming interface: TenGigabitEthernet1/1, RPF nbr 10.x.x.x, RPF-MFDOutgoing interface list:Vlan50, Forward/Sparse, 00:48:45/00:02:52, H <<====After the change:
Router# show ip mroute<<..skipped..>>(, ), 01:28:19/00:02:36, flags: sPTIncoming interface: TenGigabitEthernet1/1, RPF nbr 10.x.x.x, RPF-MFDOutgoing interface list: Null <<====Conditions: The issue is was first seen in a Cisco 7600 chassis with a Supervisor 720 engine running Cisco IOS Release 12.2(33)SRB.
Workaround: There is no workaround.
•
Symptoms: Inactive sessions are observed in the multicast explicit tracking database.
Conditions: This issue is seen when IGMP leave is received for a group followed by a CLC switchover.
Workaround: There is no workaround.
•
Symptoms: When a modem comes up during off-peak time when it is not supposed to be monitored, as per the enforce-rule configured, the SNMP is not synced to the standby line card.
Conditions: This issue is seen if a line card switchover is done during off-peak time. Even if the modem enters the peak time, iti s not monitored as per the enforced rule configured. The SNMP is synced to the standby line card once the modem is monitored during the peak time.
•
Symptoms: CMTS crashes and the following error message is seen when the no ip routing or the no router bgp xx command is run.
%IPRT-3-IPDB_DEL_ERROR: i_pdb delete error bgp, 4, 210074C8, 20E322E0, 0, 0 -Process= "IP RIB Update", ipl= 0, pid= 117, -Traceback= 0x61FD7F58 0x62005498 0x62006D24Conditions: A large number of VRFs must be configured and BGP is also configured to support these VRFs.
Workaround: There is no workaround.
•
Symptoms: The show controller modular-cable 1/0/0 command displays that rf-channel 0 becomes unknown after a SPA hw-module reset.
Conditions: This issue occurred executing the hw-module reset and show diagnostic 1/0/0 commands. This causes all wideband modems also to go online.
Workaround: Reload the router to fix the issue.
•
Symptoms: The upstream ports configured as HCCP protect show the port link down when executing the show facility-alarm status command.
Condition: This issue is seen when the cable line card is configured as HCCP protect, and the router is running a later image version than Cisco IOS Release 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: For deleted service flows associated with PCMM calls, the last character in the Service Class Name field is dropped in SAMIS records as well as in the SNMP MIB object "docsQosServiceFlowLogServiceClassName".
Conditions: This issue is seen when the dynamic service flows associated with PCMM calls are deleted. The last character is missing from the service class name in the MIB object "docsQosServiceFlowLogServiceClassName" and SAMIS records.
•
Symptoms: While executing the clear counter command, service flow statistics are not getting cleared under the wideband interface.
Conditions: This issue is seen in cases of multicast and the forwarding is through the wideband interface. Multicast service flows are created and the counters are maintained under the wideband interfaces.
Workaround: There is no workaround.
•
Symptoms: Both the working card and protect card are active.
Conditions: This issue is seen when a line card switchover happens from 5/0 to 5/1. After synchronization, shut down the 5/0/2 interface, and then revert from 5/1 to 5/0. If the 5/0/2 interface is not shut down, and after the suspend timer interval expires, both 5/0 and 5/1 are active.
Workaround: Do not shut down an active protect interface. If both the cards (working and protect) are active, reset the protect card.
•
Symptoms: Cable metering fails and enters a "hung" state.
Conditions: The issue happens when the "ip tcp timestamp" option is configured globally.
Workaround: Do not use the "ip tcp timestamp" option.
•
Symptoms: Traffic fluctuation only happens when the WB/MC topology and traffic rate are exactly the same.
Conditions: This issue is seen in Cisco IOS release 12.2SC with HQF scheduler. When 46 pecent of bandwidth is allocated to Channel 0 for MC0 and 46 percent of the bandwidth is allocated to Channel 1 for MC1,the rest of the bandwidth is allocated to WB0.
Thus, the traffic rate of WB is higher than the configured percentage. The traffic rate of MC0 is a little lower than the configured percentage and the traffic rate of MC1 is a little higher than the configured percentage.
Workaround: Do the following:
–
–
–
•
Symptoms: During line card high availability, the CM instance bpi_enable flag on the Route Processor (RP) is not synced over correctly.
Conditions: This issue is seen while generating multicast stream to CMs, and if BPI CM is used during a line card switchover.
Workaround: There is no workaround.
•
Symptoms: The MAC Domain Descriptor (MDD) change message is generated on all the RF channels globally for any RF channel configured on the CMTS.
Conditions: This issue occurs when an MDD change is supposed to be generated for channels within the MAC domain.
Workaround: There is no workaround.
•
Symptoms: When a Cisco uBR10012 router has a lot of CMs, the show cable modem docsis version summary command total may not be able to display each column properly because there is not enough space between columns.
Conditions: This issue is seen when there are a lot of cable modems. This issue is seen in Cisco IOS Release 12.3(23)BC and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: When the cable sync-interval command is configured, the cable interface needs to execute a shut or no-shut command for the interval change to take effect.
Conditions: This issue occurs while configuring the cable sync-interval command.
Workaround: There is no workaround.
•
Symptoms: The downstream Committed Information Rate (CIR) reservation for wideband cable modems works as expected, however they are displayed incorrectly under admission control command.
Conditions: This issue is seen in any w-online state wideband cable modem with non-zero CIR reservations. It displays the CIR reservation as 0 although the reservation admit or unadmit works properly. The show interface admission-control reservation downstream command displays the 0 CIR for the wideband CM.
Workaround: Use the show cable modem qos command to display the correct reservation.
•
Symptoms: When either CiscouBR-MC5X20U/S/H card MAC domains have wideband associations, a traceback may be observed along with a message similar to the following when NO CARD x/x is used:
*Jan 25 05:41:57.235 PST: %SYS-3-BADLIST_DESTROY: Removed a non-empty list(2F3F7BF4, name: (unknown)), having 36 elements -Process= "C10K Card Actions", ipl= 0, pid= 75A small amount of memory leak may also happen. No other side effects are expected.
Conditions: This traceback is seen only if the card had a NB-WB association previously, and was powered down or removed.
Workaround: There are no workarounds.
•
Symptoms: Port-channel interfaces used in conjunction with 802.1q subinterfaces show a number of problems on both the PRE4 and PRE2 processor modules. This was observed while testing the Cisco SPA-5X1GE-V2 in Cisco IOS Release 12.2(33)SCB. The problem appears to be with the 4-jacket WAN interface, and not with the older half-height Gigabit Ethernet interfaces.
Conditions: This issue occurs in a PRE2 processor module under the following conditions:
–
–
This issue occurs in a PRE4 processor module under the following conditions:
–
–
Workarounds: To restore the traffic, do the following:
1.
2.
3.
4.
Theoretically, you can boot the CMTS with no Ethernet channel configured, and copy an edited production configuration file to running-config after all the interfaces are up.
•
Symptoms: The w-online state modem comes online after protect line cards are reset before a resynchronization.The CMTS is configured with N+1 redundancy. If the working line card (7/0) is reset, the modem switches to protect (5/0). If the protect (5/0) line card is reset before resynchronization, the modem disappears, and comes in online instead of being in the w-online state.The clear cable modem reset command does not work to bring the modem back to the w-online state.
Conditions: This issue occurs while resetting an active line card during the standby line card bootup period, but before the static synchronization of an active to standby line card finishes.
Workaround: Reload the line card or the CMTS.
•
Symptoms: An IRONBUS_FAULT is displayed while pulling out the Cisco SIP-600.
Conditions: This issue is seen in Cisco IOS Release 12.2(33)SCB image running on a Cisco uBR10012 router.
Workaround: There is no workaround.
•
Symptoms: A spurious memory access was observed on a cable line card when Online Insertion and Removal (OIR) of the Cisco SIP-600 was done on a PRE 4 processor module.
Conditions: This issue is observed on PRE 4 processor module with an active wideband SPA.
Workaround: There is no workaround.
•
Symptoms: The show cable bundle multicast command does not display the entire bundle in the interface column.
Conditions: This issue is seen in the output of the show cable bundle multicast command.
•
Symptoms: Multimedia Terminal Adaptor (MTA) CMs are not accessible from the CMTS or Internet because the ARP entry is replaced with an all-zero MAC address. Hence it is not accessible when a crafted packet is sent.
Conditions: This issue occurs when a crafted packet is sent to the CMTS, causing the CMTS to use the wrong MAC address for the ARP table.
Workaround: Execute the cable source-verify dhcp server x.x.x.x command, where x.x.x.x is your DHCP server, and use cable arp filter to limit the potential ARP storm.
•
Symptoms: Wideband CMs using a stats index are not managed by the modular-host. This causes two modems using the same Blaze header stats index, thus leading to incorrect downstream traffic statistics on these modems.
Conditions: This issue occurs when there are wideband modems in the w-online state and a change is made to the downstream channel ID of the MAC domain host of these wideband modems.
Workaround: There is no workaround.
•
Symptoms: The output of "sid" is incorrect in the show cable modem x.x.x.x service-flow [verbose] command.
Conditions: This issue is seen in Cisco IOS Release 12.3(23)BC and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: Traceback and IPC error messages may appear after the CMTS is loaded with the Cisco IOS Release 12.2(33)SCB1 image.
Conditions: This issue is seen in routers running Cisco IOS Release 12.2(33)SCB1.
Workaround: There is no workaround.
•
Symptoms: The show pxf cable multicast-echo does not display the entire interface when the subinterface is greater than 13 characters.
Conditions: This issue is seen for subinterfaces greater than 13 characters.
•
Symptoms: Modifying the second-choice-channel-width has no affect on the configuration even though the cable upstream channel-width command accepts the input.
Conditions: This issue is seen in Cisco IOS Release 12.3BC.
Workaround: Modify the first-choice-channel-width, and then reconfigure the first and second choice.
•
Symptoms: The packet drop counter is not getting updated in show interface service-flow counter for single-session multicast service-flows
Conditions: This issue is seen in single session multicast service-flows.
•
Symptoms: No multicast-session is permitted when cable multicast authorization profile (mauth) is enabled and the debug messages display "session limit reached".
Conditions: This issue is seen in the following conditions:
–
–
–
–
–
Workaround: Delete the cable modem and allow the modem to come online with second cable modem configuration (non-mauth cm-file).
•
Symptoms: The CMTS may initiate DCC/DSA/DSC when Dynamic Bonding Change (DBC) is outstanding.
Conditions: If the customer is running Video over DOCSIS (VDOC) and Cisco Load Balancing (or CMTS initiated DSA) with a DPC3000, the CMTS makes VDOC higher priority over load balancing (or CMTS-initiated DSA) when a DBC is outstanding. According to DOCSIS 3.0 specifications, the CMTS must not issue a DCC or DSA/DSC command when it has previously issued a DBC command and that command is still outstanding.
Workaround: There is no workaround.
•
Symptoms: CMTS does not reject the CM-initialized DSC/DSA messages during Dynamic Bonding Change (DBC) or Dynamic Channel Change (DCC).
Conditions: This issue is seen if the customer is running Video over DOCSIS (VDOC) with a Cisco DPC3000 cable modem, and the CM is making a DSA call. The CMTS responds with a confirmation code of "reject-temporary" for the DSA-REQ or DSC-REQ for Annex B.
Workaround: There is no workaround.
•
Symptoms: After issuing the no cable trust command on the active PRE module, that same command is not removed from the standby PRE module and as a result the active and standby PRE modules are out of sync.
Conditions: This behavior is only seen if the MAC address used in the cable trust command does not exist in the CMTS.
Workaround: There is no workaround.
•
Symptom: Traffic does not resume after a PRE module switchover and modems are not accessible after a PRE module switchover.
Conditions: This issue occurs after a PRE module switchover and "cef"is not running in the standby PRE module.
Workaround: There is no workaround.
•
Symptoms: After unconfiguring the cable source-verify dhcp command, the "SrcVfy" field continues to display "On".
Conditions: This is seen when the cable ipv6 source-verify and cable source-verify dhcp commands are configured on the same bundle interface, and when the cable source-verify dhcp command is unconfigured before the cable ipv6 source-verify command.
•
Symptoms: When configuring cable upstream threshold cnr-profile snr-profile, the parser shows and accepts the second threshold value. It does not give an error message, and does not display the configuration with only the first threshold value.
Conditions: This issue is seen while configuring the CNR and SNR thresholds for modulation profiles.
Workaround: Either enter any valid threshold value or "0" to bypass.
•
Symptoms: The value range of the MIB object "ccmtrCollectionInterval" is not consistent with the command. The valid range for MIB object is 15 to 1440, but for the command, the range is 2 to 1440.
Conditions: This issue occurs while setting the MIB object "ccmtrCollectionInterval" using the SNMP set operation.
Workaround: Set the interval using the command.
•
Symptoms: A memory leak was observed after deleting the subinterface.
Conditions: This issue is seen while deleting a subinterface.
Workaround: There is no workaround.
•
Symptoms: When applying modulation profiles to an upstream port, the modulation profiles are re-ordered, high performance to robust/low performance. Modifying the order of the modulation/symbol rate on the profile does not happen.
Conditions: If the signal-to-noise ratio (SNR) threshold is set at 25 with a hysteresis of 2 (bypassing all FEC), the modulation on the upstream will be QPSK until the SNR falls below 25 dB where it will change to 16-QAM. When the SNR reaches 27 dB (25 dB + 2 dB), the modulation reverts to QPSK. When a line card switchover is performed, the modulation profiles are re-ordered on the protect card.
Workaround: Remove and re-apply modulation-profile configuration on the upstream port after modifying the modulation profiles.
•
Symptoms: All CMs on adjacent downstreams are offline. The adjacent downstream ports are 0 and 1, or 2 and 3, or 4.
Conditions: The downstream ports that share JIB may be nonfunctional if the JIB that services both downstream ports is locked.
Workaround: Reload the cable line card.
•
Symptoms: A memory leak was observed after executing the multicast test case.
Conditions: This issue is seen when the cable multicast group-qos and cable multicast qos group priority commands are unconfigured.
Workaround: First unconfigure the session-range using the cable multicast qos group priority command.
•
Symptom: Removing the line card configuration causes %LINEPROTO-5-UPDOWN and scm sum total, scm vendor sum, and scm show modem commands show no CMs.
Conditions: This is seen when line card high availability is removed and while executing scm sum total, scm vendor sum. and scm show modem commands.
Workaround: There is no workaround.
•
Symptoms: PXF forwards traffic up to 300 Mbps to the Route Processor (RP) using 1500-byte IPv6 packets.
Condition: This issue occurs when Physical Logical Bandwidth Limiting Traffic (PBLT) for the RP CPU queue is set to 300 Mbps. The show pxf cpu queue command displays this information.
Workaround: Change the value to 44 Mbps (DS3 Bandwidth) for the PRE2 processor module and 1.6 times 44 Mbps for the PRE4 processor module.
•
Symptoms: The MIB object "entPhysicalName" for theCisco uBR10012 router does not display the values correctly. The values as displayed as slot 0/0/0 and slot 0/1/0 for the PRE processor modules instead of being displayed as 0A and 0B from left to right. Similarly, the MIB object "entPhysicalName" returns slot x/x/x (for example, 1/0/0) instead of the values displayed in the device.
Conditions: This issue occurs while querying the MIB object "entPhysicalName".
Workaround: There is no workaround.
•
Symptoms: When applying new RF channels in Cisco IOS Release 12.2(33)SCB, a corresponding modular-cable interface is created. These interfaces appear under router RIP, router OSPF, and router isis routing protocol processes as "no passive-interface", even when passive-interface default is configured. Cable, modular-cable, and wideband interfaces should not appear under a routing protocol process as there is no Layer 3 control.
Condition: The creation of new cable bundle interfaces is configured rightly using the passive-interface default command. Because there is no Layer 3 control on cable, modular-cable, or wideband cable interfaces, such a problem with the Layer 2 interfaces appearing as "no passive-interface" does not lead to improper routing protocol control, but this can create some confusion even though the issue is cosmetic.
•
Symptoms: A crash is observed on the CMTS.
Conditions: This issue is seen in cases ETDB show commands are paused and multicast sessions removed because a CPE leave or a CM goes offline.
Workaround: Set the term length to 0 for no pausing while displaying the command.
•
Symptoms: The following configuration on the uBR10000 Series Universal Broadband Routers (PRE-2) can cause a system reload on both the primary and standby.
ip access-list extended CABLEINdeny udp A.0.0.0 0.255.255.255 B.C.D.0 0.0.0.255 eq tftp logdeny tcp A.0.0.0 0.255.255.255 B.C.D.0 0.0.0.255 eq 69 logpermit ip any anyWorkaround: There is no workaround.
•
Symptoms: The CMTS allows any upstream to be attached to any connector without any warning message. A "no" is displayed for the conflicting connector.
Conditions: This issue is seen when a new card is plugged into a slot with no pre-configuration.
Workaround: Reload the line card or use default connectors.
•
Symptoms: The range for Security Association Identifier (SAID) is not correct. The correct range should be 8192 to 12272.
Conditions: This issue is seen in routers running Cisco IOS Release 12.2(33)SCA.
•
Symptoms: Dynamic PCMM service flow disappears after 520U line card switchover or reset.
Conditions: This issue is observed when switchover occurs after PCMM gate built about [T3-20, T3] seconds or after sequential switchover/revertback.
Workaround: Change pcmm policy to rescue pcmm gate from COMMIT_RECOVERY back to COMMIT status.
•
Symptoms: The PRE4 processor module rate shaping allows the traffic throughput on a service flow to exceed its maximum rate.
Conditions: This issue is seen on PRE4 processor module best effort service flows.
Workaround: There is no workaround.
•
Symptoms: A crash is observed on the ESR-PRE2 processor module. The crash infomation logs may contain the following:
%UBR10K-6-US_SFID_INCONSISTENCY: US-SF found: SFID xxxx, type 0, sid 0(yyyy), MAC aaaa.aaaa.aaaa(bbbb.bbbb.bbbb), prim_sid xxx(yyy)CMD: 'no cable service attribute voice-enabled downstream-type HA-capable'Conditions: This issue is seen in Cisco IOS Release 12.3(23)BC4 and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: The show inventory command on a uBR10012 router running Cisco IOS Release 12.2(33)SCB1 may show duplicates of the DTCC cards in the chassis.
Workaround: There is no workaround.
•
Symptoms: After a reload, cable metering source-interface FastEthernet0/0/0 is not present at the running configuration, although cable metering filesystem disk0: cpe-list-suppress and cable metering source-interface FastEthernet0/0/0 were configured.
Conditions: This issue is seen only in Cisco IOS release versions on the PRE2 processor module and not in Cisco IOS releases versions of the Cisco uBR7200 router.
Workaround: There is no workaround.
•
Symptoms: The active DOCSIS queues remain active in the event of a DOCSIS SPA failure.
Conditions: If there are modems online and passing traffic on the modular-cable and/or on wideband-cable interfaces and the DOCSIS SPA fails, the DOCSIS queues associated with these online modems remain active attempting to dequeue.
Workaround: There is no workaround.
•
Symptom: When attempting to upgrade a Cisco uBR10012 router from Cisco IOS Release 12.3(21)BC to Cisco IOS Release 12.2(33)SCB, the primary PRE may crash.
Conditions: This issue is seen when the CMTS is running Cisco IOS Release 12.3(21)BC and while upgrading to Cisco IOS Release 12.2(33)SCB. The CMTS must then be powered down and physically reconfigured to run as a M-CMTS (Cisco SIP-600 inserted, WB SPAs inserted, DTCCs inserted, so on.). Lastly, the CMTS must be turned on to run its new Cisco IOS Release 12.2(33)SCB IOS image and physical configuration.
•
Symptoms: The RFGW-1 reports channel(s) as down following rf-channel and fiber node configuration changes.
Conditions: This issue is seen when there are changes made to the rf-channel/fiber node configuration on the RFGW-1. It reports the affected channels as DOWN.
Workaround: Exit fiber node configuration level and the changes are be committed and the rf-channel status as reported by the RFGW-1. It shows the correct state.
•
Symptoms: Modems marked by tftp-enforce mark-only show up in show cable modem as marked with the # sign. Syslog messages might report that the modem was marked as follows:
SLOT 6/1: Mar 25 12:21:58.967 CDT: %UBR10000-4-REGISTRATION_BEFORE_TFTP_MARK: Registration request unexpected: Cable Modem did not attempt TFTP. Modem marked with #. CM Mac AddrBut the modem does not show up with the "#" in front of the online status in show cable modem command.
Conditions: This is seen in routers running Cisco IOS Release 12.2(33)SCB1.
Workaround: There is no workaround.
•
Symptoms: The show facility-alarm status command displays a false alarm.
Conditions: This issue is seen for the protect line cards when connectors are not set.
Workaround: There is no workaround.
•
Symptoms: When a bundle subinterface number is greater than 64K, the statindex assignment for multicast session shows 0.
Conditions: This is a rare situation and only happens when a subinterface number is greater than 64K.
Workaround: Use a smaller subinterface number.
•
Symptoms: The CMTS sends DBC to the CM to change the primary downstream port
Conditions: This issue is seen when the VDOC broadcast feature is configured such that multiple RCCs in the same MAC domain have exact same frequencies for the primary bonding group, but the primary frequency is different. For example, if you have frequencies f1 and f2 as the primary bonding group frequencies, then rcc1 (f1f2f3) and rcc2 (f1f2f4) have the first two frequencies as f1f2, while rcc3 (f2f1f3) and rcc4 (f2f1f4) have the first two frequencies as f2f1. With this configuration, the CM is first online using rcc3 and the CPE sends an IGMP join for the channel corresponding to f3.
Workaround: For the VDOC broadcast feature, configure RCC templates such that a set of primary bonding group frequencies are in the same order in all RCC templates.
•
Symptoms: Executing the cmts walk command on ARP table causes high CPU utilization. This is also seen while using the snmp walk command on the ARP table.
Conditions: This issue occurs while using the cmts walk and snmp walk commands.
Workaround: To prevent high CPU utilization when using the snmp walk command, use snmp view commands:
snmp-server view cutdown iso included
snmp-server view cutdown at excluded
snmp-server community public view cutdown ro
snmp-server community private view cutdown rw
•
Symptoms: Although the CM is in non-BPI mode, after line card high availability, its downstream packet or the ping are still encrypted and fail.
Conditions: This issue is seen in the following conditions:
–
–
–
Workaround: After the change to non-BPI configuration, use the delete command instead of reset.
•
Symptom: The CM goes offline after line card fails with different types of working and protect line card (for example, Cisco uBR10-MC5X20H protect and Cisco uBR10-MC5X20U line cards)
Conditions: This issue occurs when a manual line card switchover is done and when a new active line card crashes or reboots before HCCP Static Sync is done.
Workaround: There is no workaround.
•
Symptoms: The show interface bundle command output displays incorrect upstream input bit/sec traffic counter.
Workaround: There is no workaround.
•
Symptoms: After a line card switchover the key-index becomes 0 for a multicast QoS that is configured for a active static IGMP session along with encryption.
Conditions: This issue is seen only for static IGMP groups which has multicast QoS with encryption configured.
Workaround: There is no workaround.
•
Symptoms: An unused Downstream Service Identifier (DSID) and statsindex is shown in the show cable multicast dsid command. The Service Flow Identifier (SFID) value displays "0" in the show interface multicast-session command.
Conditions: This happens when admission control checks fails at the cable line card (CLC) during multicast service flow creation.
Workaround: There is no known workaround.
•
Symptoms: Cable modems report non-primary DS channel impairments during registration time via the partial-service in the REG-ACK message.
Workaround: There is no workaround.
•
Symptoms: A uBR10012 router may see the following syslog generated repeatedly:
%UBR10000-6-MAP_UPSTREAM_BW_ALLOCATION: A transmit opportunity was missed because MAP arrived too lateWorkaround: There is no known workaround.
•
Symptoms: Service flow statistics wrongly resets to 0 after executing the dcc technique 1 command.
Workaround: There is no known workaround.
•
Symptoms: The Gigabit Ethernet interface on an ESR-1GE may fail to send packets.
Conditions: This might be seen just after executing online insertion and removal (OIR) of an ESR-1GE with low rate.
Workaround: Execute the hw-module slot reset command.
•
Symptoms: The multicast packets are forwarded to a cable downstream interface that is in rf-shutdown state.
Conditions: This issue occurs if the cable interface is in rf-shutdown state and multicast traffic is forwarded to that interface unnecessarily.
Workaround: There is no workaround.
•
Symptoms: Multicast traffic forwarding stops on modular cable interfaces.
Conditions: This issue occurs on modular downstream interfaces with multicast QoS functionality configured and the Parallel Express Forwarding (PXF) on the CMTS has restarted.
Workaround: In case of a static multicast, unconfigure and reconfigure the ip igmp static-group commands. In case of dynamic multicast, use the clear ip mroute command.
Workaround: There is no workaround.
•
Symptoms: When using the RCC template with multiple channels marked as primary, the cable modem will see sporadic downstream traffic loss depending on the modem's selection of the primary downstream.
Conditions: The cable modem must be locked on to the primary downstream that is not RC-1 in the RCC template.
Workaround: Assign only a single RC/DS as primary per RCC-template.
•
Symptoms: Index allocation might fail due to an IPC drop between the host and the working line card, or because the guardian is down.
Workaround: Reset the line card, guardian, and the SPA.
•
Symptom: Logged messages are sent not only to the IOS system log in memory on the Cisco MC5X20 line cards, but to the console serial port. This I/O activity uses CPU resources unnecessarily especially at card bootup time. Console logging may be re-enabled by configuring the logging console command.
Conditions: This issue is seen in all Cisco uBR10-MC5X20 line cards.
•
Symptoms: The Cisco uBR10012 router crashed because it ran out of unfragmented memory.
Conditions: This is seen in Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: Multicast session is in I-state on the active PRE.
Workaround: There is no workaround.
•
Symptoms: In Cisco IOS Release 2.2(33)SCB3, executing the show controller modular-cable x/y/z command with a copper SFP, SFP-GE-T, causes the wideband SIP to crash.
Conditions: The issue will occur if we have Copper SFP in a Wideband SPA gigabit ethernet port.
Workaround: Use a fiber SFP and a fiber link instead of a copper link.
•
Symptoms: Cable modems do not respond to layer 3 pings.
Conditions: This issue is only on the Cisco uBR10-MC5X20S and Cisco uBR10-MC5X20U cards and when PRE-EQ (equalization-coefficient) is configured.
Workaround: Restart the cable line card.
•
Symptoms: A multicast QoS session is not automatically created when changing group configuration or GQC.
Condition: This issue is seen when a multicast QoS group is configured on a bundle interface and a matching static-group is defined. However, any change in the group configuration after it is applied to the bundle interface is not taken into effect, and the original multicast QoS session is removed.
Workaround: Remove the multicast QoS session from the bundle interface and re-apply it to the bundle.
•
Symptoms: Alignment traceback messages are found when the show memory leak res user" and show buffer usage commands are executed.
Workaround: There is no workaround.
•
Symptoms: Potential wideband SPA Index leak is found during the line card switchover.
Conditions: This issue is found when the modem is deleted before the IPC is returned from the guardian.
Workaround: There is no workaround.
•
Symptoms: The docsIfUpChannelTxBackoffStart/end value is inconsistent with the output from the command value.
Workaround: There is no workaround.
•
Symptoms: During the VDOC video channel changes, it is possible that the multicast QoS service flow cannot be created because a stale multicast QoS service flow exists for 140 seconds after the last IGMP member. This prevents a new multicast QoS service flow from being created.
Conditions: This is seen in multicast QoS service flows.
•
Symptoms: The "Gate Report State" counter in the output of the show packetcable cms verbose command does not increment for PacketCable Multimedia (PCMM) policy servers.
Conditions: This issue is seen for PCMM policy servers.
Workaround: There is no workaround.
•
Symptoms: Memory leak is observed due to SID management task.
Conditions: This issue is observed when the packetcable calls are present.
Workaround: There is no workaround.
•
Symptoms: Memory leak is observed on the linecard due to MAC parser with overnight PC calls.
Workaround: There is no workaround.
•
Symptoms: HQF logging does not work on Cisco PRE4.
Workaround: There is no workaround.
•
Symptom: CMTS will accept inconsistent "per-SF L2VPN TLV" in the cable modem configuration file.
Conditions: This is seen in the following conditions:
–
–
–
Workaround: There is no workaround.
•
Symptoms: CMTS may crash when an MPLS sub interface is deleted.
Workaround: Do not delete the sub interface.
•
Symptoms: OIR of a PCMCIA flash disk can cause a watchdog reset of the PRE.
Conditions: This issue is observed during OIR of the PCMCIA flash disk that is being actively accessed.
Workaround: Restrict the removal of PCMCIA flash disks.
•
Symptoms: CM uptime mismatch is reported between the CMTS and few CMs.
Workaround: There is no workaround.
•
Symptoms: CPU increases to 100% on the Cisco MC5x20U and MC5x20S line cards after upgrading to the Cisco IOS Release 12.2.(33)SCB2.
Workaround: Downgrade the image version to Cisco IOS Release 12.3(23)BC7 or use MC5x20H cards.
•
Symptoms: Legitimate CM which was online and BPI enabled will get disconnected along with the following error messages printed with "CLONED_CM_DETECTED" corresponding to legitimate modem's MAC address.
Workaround: Once the legitimate CM goes offline and comes back again online with BPI enabled, the next cloned CM attempt will fail.
•
Symptoms: In a SAMIS operation, counters are stuck in the downstream direction for cable modems.
Workaround: Reset the cable modems.
•
Symptoms: Traceback is seen at cable_hwidb2macdomain.
Workaround: Do not pre-configure the card command with ubr10k-clc-5x20h. The card command to use for the MC520H cards is card x/x 5cable-mc520h-d.
•
Symptoms: Tracebacks are seen on the Cisco uBR10012 router after two multicast clients initiate multiple joins and leaves.
Workaround: There is no workaround.
•
Symptoms: Entries in docsBpi2CmtsMulticastObjects duplicate after the line card revertback.
Workaround: There is no workaround.
•
Symptoms: The modems on a specific line card cannot come online after upgrading to Cisco IOS Release 12.3(23)BC4.
Conditions: The following message is seen in the logs report the message:
%PXF_DMA-3-IRONBUS_NOTRUNNING: Data path to slot 5/0 failed to synchronize.Workaround: There is no workaround.
•
Symptoms: Spurious memory access occurs or CMTS crashes.
Conditions: This issue is observed when DSG configuration is removed and protect linecard is active.
Workaround: Do not update configuration when protect linecard is active.
•
Symptoms: Multicast functionality stops when the modems are reset on a different interface.
Workaround: Change to different multicast stream and then change back to original stream to restore the stream.
•
Symptoms: Wrong DS frequencies are offered to wideband CMs.
Workaround: There is no workaround.
•
Symptoms: CPU utilization shoots to 100% for multiple times during FPGA upgrade.
Workaround: There is no workaround.
•
Symptoms: Non-bonding capable modems are wrongly registered in bonded channels.
Workaround: There is no workaround.
•
Symptoms: Improper SPA temperature minor warning reminder message is observed.
Conditions: This issue is observed when the WBCMTS SPA temperature crosses over minor warning threshod 58C, and remains in the range of 58C ~ 68C for more than 10 minutes.
Workaround: There is no workaround.
•
Symptoms: CIR queue or service flow is not created for the multicast session.
Workaround: Run the shut/no shut interface command and the session will be recreated.
•
Symptoms: During the IPDR process, it is unable to delete config for cable metering.
Workaround: There is no workaround.
•
Symptoms: CMTS may crash due to stack corruption.
Workaround: There is no workaround.
•
Symptoms: CMTS may crash after a key-string is entered.
Workaround: There is no workaround.
•
Symptoms: Active PRE fails over to the secondary reporting:
ACTIVE RP WATCHDOG TIMEOUT TRIGGEREDConditions: This issue is observed on Cisco IOS Release 12.3(23)BC7.
Workaround: There is no workaround.
•
Symptoms: When setting cable upstream docsis mode by SNMP, the standyby PRE does not synchronize this configuration.
Workaround: Set cable upstream docsis mode by CLI instead of SNMP.
•
Symptoms: The arp entries are lost after a redundancy switchover or a crash
Workaround: There is no workaround.
•
Symptoms: The Cisco SIP-600 crashed after executing the hw-module bay 1/0 shutdown and the hw-module bay 1/1 shutdown commands at the same time.
Workaround: There is no workaround.
•
Symptoms: Tracebacks appears while unconfiguring cable service attribute non-ds-bonded downstream-type bonding-disabled and cable fiber node.
Workaround: There is no workaround.
•
Symptoms: The show interface multicast-session command may show wrong multicast session instances.
Workaround: There is no workaround.
•
Symptoms: CMTS does not add IP address to the relayed DHCP packets.
Workaround: There is no workaround.
•
Symptoms: The following CPUHOG message is seen when CPE is rapidly performing hundreds of multicast joins and leaves and a hw-module reset is done on the line card.
005812: Jul 2 11:53:06.925: %SYS-3-CPUHOG: Task is running for (2000)msecs, more than (2000)msecs (66/36),process = Net Background.-Traceback= 4188E9F8 4187AFAC 4187E2C0 4187EB70 4187F5CC 4187F860 418848D4 41884988 41884A3C 402ACBC8 40264654 40264ABC 403E87FC 403E8AF0 403E8B44 402654D8Workaround: There is no workaround.
•
Symptoms: After sending multicast traffic and then a PRE switchover, the following unexpected messages are displayed on the Standby PRE console.
Sid : 22, Interface: Cable6/1/0000715: Jul 2 22:56:44.605: CMInstp is NULL![cmts_mcast_add_session_interface]Sid : 93, Interface: Cable5/0/0000716: Jul 2 22:56:44.609: CMInstp is NULL![cmts_mcast_add_session_interface]Workaround: There is no workaround.
•
Symptoms: Non default service flows and DSID entries were created when default multicast service flow failed.
Conditions: This happens only when MQoS enabled scenario where admission control failed to create multicast default SF.
Workaround: Make valid admission control parameters for the default service class template.
•
Symptoms: A cable modem is wrongly assigned with a DS-SG-ID which does not exist on RP after the line card switchover.
Workaround: There is no workaround.
•
Symptoms: The PRE crashes when an Etherchannel subinterface with dot1q encapsultation is deleted.
Workaround: There is no workaround.
•
Symptoms: DSG traffic is no longer forwarded to modular interface after changing bundle.
Workaround: Remove and add the bundle again
•
Symptoms: Intermittent traceback message at OIR of PRE-4 flash disk.
Workaround: There is no workaround.
•
Symptoms: When changing the service class on the fly for DSG tunnel, CMTS removes the old service flow but fails to repopulate new service flows.
Workaround: Remove the DSG tunnel from the interface and re-config it.
•
Symptoms: On a uBR10012 router with Cisco SIP-600 and wideband SPA, SIP-600 went down and came back up due to the Software Watchdog Reset.
Workaround: There is no workaround.
•
Symptoms: After a crash or a OIR of a cable line card, the RCC template is not applied to cable interface.
Conditions: LCHA must be enabled. The initially active card crashes and comes up and becomes active.
Workaround: When in this state unconfigure and re-configure the associated fiber-node.
•
Symptoms: Active PRE4 reloaded after the hw-module bay 1/1/0 reload command was run.
Workaround: There is no workaround.
•
Symptoms: Traceback messages have been observed while executing the show interface command.
Workaround: There is no workaround.
•
Symptoms: Traceback messages have been observed after pulling the active PRE-B (PRE4) while it was active.
Workaround: There is no workaround.
•
Symptoms: The MC5x1GE WAN SPA crashed after OIR.
Workaround: There is no workaround.
•
Symptoms: CMTS watchdog timeout crash is observed at cmts_address_filter or cmts_cm_lookup.
Workaround: Downgrade the image version to Cisco IOS Release 12.2(33)SCB2.
•
Symptoms: Memory fragmentation might occur with memory allocation of 65535 bundle MAC entry elements.
Conditions: This issue is likely to occur with multicast traffic.
Workaround: There is no workaround.
•
Symptoms: Some cable modems may go offline during a line card switchover process.
Conditions: This issue is observed when a Cisco MC5x20H line card that serves as a protect card and Cisco MC5x20U line card is the working card, switches over from MC5x20H line card to MC5x20U line card.
Workaround: Reduce the number of modems connected to the Cisco MC5x20U line card.
Resolved Caveats — Cisco IOS Release 12.2(33)SCB4
•
Symptoms: The no ip dhcp relay info policy removal pad config command has no effect.
Workaround:
1.
2.
3.
•
Symptoms: Memory leak at CMTS_HCCP_GET_TLV_PARSE_TLV has been observed at the RP.
Workaround: There is no workaround.
•
Symptoms: It was not possible to ping the wideband cable modem after the test cable dcc tech 1 command is executed.
Conditions: This issue is observed when the cable modem is in the w-online MAC state.
Workaround: Use tech 0 for DCC operation for wideband cable modems.
•
Symptoms: A cable line card crash may occur if a test cable dcc command is executed on a 2.0 modem, move it from an ATDMA upstream to a TDMA upstream (with fragmentation disabled).
Condition: The issue is seen in routers running Cisco IOS Release 12.3(23)BC4.
Workaround: There is no workaround.
•
Symptom: Invalid/ corrupt values seen for OctetsPassed and PacketsPassed fields in SAMIS records.
Conditions: CMTS with wideband SPA configured, query the service flow counters using SAMIS, snmp or show CLI, sometimes it will show wrong value for the octets or/and packets for the wideband modems or narrowband modems but using the remote SPA channel as primary downstream.
Workaround: There is no workaround.
•
Symptoms: The startup configuration information for a standby PRE is not displayed.
Workaround: There is no workaround.
•
Symptoms: CMTS crashes after performing multiple IGMP join and leave messages.
Conditions: This issue is observed when multiple IGMP join and leave messages are transmitted and with a standby PRE available.
Workaround: There is no workaround.
•
Symptoms: After a PRE processor module switchover, the following message is displayed:
"*Jan 29 10:22:33.187: %IPC-5-INVALID: Message Port index Dest Port 0x1001C Session 0x1 Source 0x60000"Conditions: This issue has been observed on In Service Software Upgrade (ISSU) while upgrading PRE A slot to PRE B slot processor modules.
Workaround: There is no workaround.
•
Symptoms: The Cisco Broadband Troubleshooter trace window appears to show incorrect data when triggering an upstream by CM MAC address. The MIB object "server% getmany -v2c ccsSpectrumDataPower" returns values that when graphed do not show the expected QAM haystack. This MIB does not provide data until the CMTS is first configured appropriately via SNMP sets.
Conditions: This issue occurs in upstream frequency stacking and configured MAC domains.
Workaround: Remove the frequency stacking and configured MAC domains.
•
Cisco devices running affected versions of Cisco IOS Software are vulnerable to a denial of service (DoS) attack if configured for IP tunnels and Cisco Express Forwarding.
Cisco has released free software updates that address this vulnerability.
This advisory is posted at http://www.cisco.com/warp/public/707/cisco-sa-20090923-tunnels.shtml.
•
Symptom: The title field "Name" in the show cable admission-control interface is missing.
Conditions: This is seen in the output of the show cable admission-control interface command.
Workaround: There is no workaround.
•
Symptoms: The wideband cable interface is not able to pass traffic.
Conditions: The issue occurs because the bonded channel-to-RF channel mapping configuration is missing from the SPA.
Workaround: There is no known workaround.
•
Cisco IOS devices that are configured for Internet Key Exchange (IKE) protocol and certificate based authentication are vulnerable to a resource exhaustion attack. Successful exploitation of this vulnerability may result in the allocation of all available Phase 1 security associations (SA) and prevent the establishment of new IPsec sessions.
Cisco has released free software updates that address this vulnerability.
This advisory is posted at http://www.cisco.com/warp/public/707/cisco-sa-20090923-ipsec.shtml
•
Symptoms: Upstream-and-downstream-related tables are not retrieved while querying via SNMP.
Conditions: This issue is seen while querying via SNMP.
Workaround: There is no workaround.
•
Cisco IOS Software configured with Authentication Proxy for HTTP(S), Web Authentication or the consent feature, contains a vulnerability that may allow an unauthenticated session to bypass the authentication proxy server or bypass the consent webpage.
There are no workarounds that mitigate this vulnerability.
This advisory is posted at the following link:
http://www.cisco.com/warp/public/707/cisco-sa-20090923-auth-proxy.shtml
•
Symptom: The Cisco uBR10012 router experienced several crashes when upgrading from Cisco IOS Release12.3(21a)BC6 to Cisco IOS Release 12.3(23)BC6.
Conditions: These crashes have been found when upgrading from Cisco IOS Release 12.3(21a)BC6 to Cisco IOS Release 12.3(23)BC6.
Workaround: There is no workaround.
•
Symptoms: The PRE crashes when the no cable modem remote-query command is run.
Conditions: This issue may occur only when the minimum polling interval is "cable modem remote-query 1 public".
Workaround: The polling interval should not be a minimum value.
•
Symptoms: Different values for the MAC Domain Downstream Service Group (MD-DS-SG) are seen after an RPR switchover, using the show cable modem wideband and show cable mac-domain downstream-service-group commands.
Conditions: This issue occurs while changing the MD-DS-SG by changing the fiber node when CMTS is up and after a RPR switchover.
Workaround: There is no workaround. This symptom is not critical. It does not affect normal use of relative features.
•
Symptoms: The show controller modular-cable command output displays invalid voltage measurement readings for the 24 rf-channel SPA.
Conditions: This is seen on the 24 rf-channel SPA.
Workaround: Re-execute the show controller modular-cable command.
•
Symptoms: When attempting to upgrade a Cisco uBR10012 router from Cisco IOS Release 12.3(21)BC to Cisco IOS Release 12.2(33)SCB, the secondary PRE processor module may crash directly after the primary PRE processor module crashes.
Conditions: This issue occurs in routers running Cisco IOS Release 12.3(21)BC.
•
Symptoms: Due to the way the CMTS US scheduler is designed, the VoIP traffic from a single CMTS shows the tendency to get synchronized in time.
Conditions: This issue is seen when multiple CMTSs are connected to the same DTI server, and the egress VoIP traffic from these CMTSs to get synchronized. This multi-CMTS VOIP synchronization causes the peak VoIP rate on the aggregation routers to the levels which default buffers cannot be handled.
Workaround: If the UGS scheduling across US channels is randomized on a single CMTS, it may to resolve the issue.
•
Symptoms: Facility alarms are incorrectly generated for cable line card interfaces.
Conditions: This issue occurs when line card redundancy is configured or upstream frequency stacking is configured.
Workaround: There is no workaround.
•
Symptoms: Cable modems report non-primary DS channel impairments during registration time via the partial-service in the REG-ACK message.
Workaround: There is no workaround.
•
Symptoms: If a modular-cable channel with online modems is removed from its host and reconfigured to another host, the flows counter are incorrectly displayed.
Conditions: This is seen when a modular-cable channel is removed from its host and reconfigured to another host.
Workaround: There is no workaround.
•
Symptoms: During the PRE switchover, Subscriber Account Management Interface Specification (SAMIS) feature may not successfully export data.
Workaround: There is no workaround.
•
Symptoms: The fiber node accepts non-existent upstream connectors, which may confuse the user on physical and logical topology.
Conditions: This issue occurs when configuring non-existent upstream connectors on the fiber node.
Workaround: Do not configure non-existent upstream connectors for the fiber node.
•
Symptoms: A cable modem with multiple service flows failed to respond to pings from the RP after the line card switchover.
Workaround: There is no workaround.
•
Symptom: The DOCSIS Timing, Communication and Control (DTCC) card information is displayed in the output of the show diagnostic command when the card is not present.
Conditions: This issue is seen when both SPA cards are present in the slot/subslot and utility cards are present in slot/subslot. For example, a SPA is present in 1/1 and a DTCC card is present in 1/1.
Workaround: There is no workaround.
•
Symptom: The number of wideband modems collected via snmp walk and show cable modem wideband command does not match.
Conditions: This is an intermittent problem and is not seen on all routers.
Workaround: Run the show cable modem wideband command to know the exact number of wideband modems.
•
Symptom: The running configuration shows "-2147483" for ip rsvp dsbm non-resv-send-limit [ burst | peak | rate ] 2147483 command, which may indicate that the CMs are not working properly.
Conditions: This issue is seen while configuring the upper limit value for ip rsvp dsbm non-resv-send-limit command.
Workaround: Do not configure the upper limit value for ip rsvp dsbm non-resv-send-limit command.
•
Symptoms: The guardian line card does not synchronize with the new key to the standby line card, when a key is renewed for a wideband multicast sid,. This causes the wrong key to be sent to a wideband CM for that multicast sid after the guardian lie card switches over.
The information is not synchronized to the standby line card when the guardian line card assigns an index to a modular-cable multicast sid and key. It may reassign, the same key after a switchover.
If the index of a remote guardian line card is not rightly made free for a modular-cable multicast sid and key, This could cause an index leak for the remote guardian line card and the information is not synced to the standby for both local and remote line cards.An index leak may happen if the guardian line card switches over.
When an modular-cable interface is shut or removed from a Channel Grouping Domain (CGD), indexes are not cleaned up on standby guardian line card causing an index leak if the guardian switches over.
Condition: This is seen after a guardian line card switchover.
Workaround: There is no workaround.
•
Symptom: The test cable dcc frequency command moves one modem to target frequency does not work when the upstream channel ID of the modem does not belong to the target downstream channel.
Conditions: This issue only affect test cable dcc frequency command.
Workaround: Use test cable dcc frequency command to move the modem where the upstream channel ID belongs to the target downstream channel.
•
Symptom: The cdxIfUpChannelAvgUtil reports incorrect numbers when rate adapt is enabled on router
Conditions: This is seen when using SNMP to poll cdxIfUpChannelAvgUtil with rate adapt enabled.
Workaround:
1) Use CLI to obtain numbers
2) Disable rate-adapt
•
Symptom: The Multimedia Terminal Adapter (MTA) lacks IP connectivity and cannot be reached.
Conditions: This is seen when upgrading to Cisco IOS Release 12.2(33)SCB2.
Workaround: Reset the MTA.
•
Symptoms: The Cisco CMTS cable line card crashes under some rare conditions when a cable modem is doing initial ranging.
Conditions: This issue may occur when an online modem goes offline and does an initial ranging before the CMTS can detect that the modem has gone offline.
Workaround: There is no workaround.
•
Symptoms: The cable dynamic-secret lock command does not lock the rogue modem to 10kbps upstream or 10kbps upstream /downstream profile. The show cable modem <mac of rogue modem> qos command reports 0/0 kbps upstream/ downstream rates for the same modem after the same modem reboots or resets.
Conditions: The issue has been observed on Cisco uBR10012 router with PRE2 processor module running Cisco 12.3(23)BC4 with cable dynamic-secret lock command configured under the cable interface.
Workaround: There is no workaround.
•
Symptom: The "PXF DMA Cumulative Ironbus Error Statistics" values are incriminated when executing the show pxf dma counters | include Ironbus command. For example, the error statistics displayed are:
PXF DMA Cumulative Ironbus Error Statistics:Ironbus Event 1/0: 0Ironbus Event 1/1: 166Ironbus Event 2/0: 2843674Ironbus Event 2/1: 2843674Ironbus Event 3/0: 70Ironbus Event 3/1: 67Ironbus Event 4/0: 63Ironbus Event 4/1: 64Ironbus Event 5/0: 241Ironbus Event 5/1: 2843674Ironbus Event 6/0: 278Ironbus Event 6/1: 268Ironbus Event 7/0: 250Ironbus Event 7/1: 269Ironbus Event 8/0: 2843674Ironbus Event 8/1: 1580072Conditions: The above errors are observed on a Cisco uBR10012 router with a PRE2 processor module running Cisco IOS Release 12.2(33)SCB1 and Cisco IOS Release 12.2(33)SCB2.
Workaround: There is no workaround.
•
Symptoms: The Arris Embedded Multimedia Terminal Adapter (EMTA) modems do not get a dial tone after a PXF crash.
Conditions: This issue is seen on a Cisco uBR10012 router running Cisco IOS Release 12.2(33)SCB2. This is seen after a PXF crash and when the Arris modems are in the "In Progress Telephony-DHCP" state.
Workaround: Switch over to the standby PRE module and execute the clear cable modem command.
•
Symptoms: The following SPA bus error and Jacket watchdog reset is displayed on a Cisco Wideband SIP crashinfo:
SPI FPGA: TIB SPA1 Bus Error [1/16]Machine Check Error, can be ECC or Watchdog..ECC 1 bit errors since last time we cleared = 0ECC 1 bit errors while up (total) = 0Conditions: This issue is seen when Cisco Wideband SPA power is shutdown and if power supply violation is watched by power management hardware to protect components on SPA.
Workaround: There is no workaround.
•
Symptom: The active time is inaccurate both downstream and upstream service flows with Cisco uBR10012 router The show interface cable x/y/c sid command indicates a correct age value.
Conditions: This issue was observed on both Cisco uBR7246VXR and Cisco uBR10012 routers running Cisco IOS Release 12.3(23)BC7 when DOCSIS QoS calls were created on a non DOCSIS QoS packetCable environment.
Workaround: Use the show interfaces cable x/yz sid command to see the correct value.
•
Symptoms: The PacketCable Multimedia (PCMM) calls fail as the Gate Set Ack/Err is not received.
Conditions: This issue is seen in PCMM calls with a small value used for timer T1 (such as 1 second).
Workaround: Use larger values for the PCMM timer T1 (at least 5 seconds).
•
Symptoms: A a long latency for VDOC video channel change is seen as the Dynamic Bonding Change (DBC) timer restricts the CMTS to have another transaction on the same CM.
Conditions: This is seen due to the DBC timer value.
Workaround: Lower the T10 timer to 10 ms.
•
Symptom: The Policy Based Routing (PBR) functionality (set ip next-hop and set ip default next-hop) does not work with the PRE4 processor module.
Conditions: This is seen in the PRE4 processor module.
Workaround: There is no workaround.
•
Symptoms: When one Cisco SIP-600 card is removed, the downstream data on the second SIP-600 card may be permanently disrupted on some service flows.
Workaround: Choose one of the following options as a workaround to this issue.
–
–
•
Symptoms: When downstream load balancing (modem count method) is configured on a MAC Domain, where DOCSIS 3.0 and DOCSIS 2.0 are mixed, the total number of CMs to be load-balanced across the downstream interfaces are incorrect.
Conditions: The wideband modem may be corrupted. The downstream interface may go into the initial state with active CMs.
Workaround: Reload the CMTS.
•
Symptoms: The "No map buffer" error is seen during a PXF crash and modem drop offline.
Workaround: There is no workaround.
•
Symptoms: The wideband cable modem goes offline after the working line card is changed during the line card switchover.
Workaround: Do not change the working line card configuration after restarting the modem.
•
Symptoms: The show inventory command output does not display the SPA's correct serial number after swapping any two SPAs.
Workaround: There is no workaround.
•
Symptoms: Admission control checking in the dynamic load balancing state may not work in some circumstances.
Workaround: There is no workaround.
•
Symptoms: Traceback is observed after running the show diag command.
Conditions: This issue was observed on the Cisco uBR10012 routers running Cisco IOS Release 12.2(33)SCB3 with WBCMTS SPA.
Workaround: There is no workaround.
•
Symptoms: After squeezing (removing the deleted files from the bootflash) the bootflash on a cable line card, the following system log is generated that may encourage reformatting of line card bootflash.
cisco#squeeze clc5_0-bootflash:All deleted files will be removed. Continue? [confirm]Squeeze operation may take a while. Continue? [confirm]Squeeze of clc5_0-bootflash completecisco#SLOT 5/0: Jun 1 14:05:24.541 CDT: %FILESYS-5-MISMATCH: bootflash device programming algorithm mismatch detected, reformat is recommendedConditions: This issue is observed on the Cisco uBR10-MC5X20S, Cisco uBR10-MC5X20H, and Cisco uBR10-MC5X20U cards running Cisco IOS Release 12.2(33)SCB2.
Workaround: Format the bootflash as recommended.
•
Symptoms: A PRE with a faulty ethernet controller for the IPC data path may indirectly cause cable line cards to crash.
Conditions: The router must be configured to use remote SPA channels using the downstream Modular-Cable command.
Workaround: There is no workaround.
•
Symptoms: More than one classifier cannot be added to the DOCSIS Set-Top Gateway (DSG) tunnel and it remains in disabled state.
Conditions: The router had the DSG source specific tunnel-group configuration.
Workaround: There is no workaround.
•
Symptoms: The interfaces using frequency stacking may get dropped up to 10 dB in the average MER readings of show controllers cable command.
Conditions: This is seen in routers running Cisco IOS Release 12.3(23)BC7.
Workaround: Remove the frequency stacking or remove the frequency on the unused upstream.
•
Symptoms: PXF crashes with the following error message in the PXF Crashinfo report.
PXF DMA Error - Invalid Source Descriptor Based on CommandConditions: This is seen in routers running Cisco IOS Release 12.2(33)SCB2.
Workaround: There is no workaround.
•
Symptoms: Packetcable gates are stuck in the ALLOC state after HW-Module reset.
Workaround: There is no workaround.
•
Symptoms: DOCSIS Set-Top Gateway (DSG) statistics does not show the modular cable interfaces Downstream Channel Descriptor (DCD) counters.
Conditions: This occurs in normal conditions.
Workaround: There is no workaround.
•
Symptom: When a modem is load balanced with Dynamic Channel Change (DCC), receives a "cdxCmtsCmOnOffNotification" OFF trap from the cmts, but no corresponding ON trap.
Conditions: This is seen on a Cisco uBR10012 router running Cisco IOS Release12.3(23)BC2 with snmp-server traps enable cable cm-onoff command configured in the global configuration and cable enable-trap [cm onoff-notification] command on each interface along with load balancing
Workaround: There is no workaround.
•
Symptoms: FPGA upgrade may fail and corrupt the SPAs that are plugged into the Cisco SIP-600.
Conditions: This issue is observed during OIR of one SPA while the FPGA of another SPA is being upgraded or hardware/ low-level software failures on the jacket card.
Workaround: Perform the following tasks.
1.
2.
•
Symptoms: A Cisco uBR10012 router running Cisco IOS Release 12.2(33)SCB2 may have a counter issue as observed with show interface command.
Workaround: There is no workaround.
•
Symptoms: CMTS may crash when the IPDR Exporter/Collector session restarts too fast.
Conditions: This issue is observed when the IPDR session is configured and the IPDR exporter is running.
Workaround: The collector does not restart the session at a faster rate, so delay the time interval to connect of 1 minute before restart.
•
Symptoms: The default facility-alarm command does not set the facility alarm values to the default values but wrongly executes the no facility-alarm command.
Workaround: There is no workaround.
•
Symptoms: Modems do not go online during CM steering even if load balancing is configured on them.
Workaround: There is no workaround.
•
Symptoms: Some signal-to-noise ratio (SNR) difference is observed when frequency stacking is configured.
Conditions: This issue is only on the Cisco uBR10-MC5X20S and Cisco uBR10-MC5X20U cards and when PRE-EQ (equalization-coefficient) is configured.
Workaround: Run the no-shut the first upstream command.
•
Symptoms: After a PRE switchover, local CMTSs can ping the affected modems but the remote hosts cannot ping modems.
Workaround: Reset the modems.
•
Symptoms: PRE and line cards crash while accessing invalid memory area.
Conditions: This issue is observed when default multicast service flow is being deleted.
Workaround: There is no workaround.
•
Symptoms: Dropped or delayed DOCSIS MAC management packets resulted in problems such as call failures.
Conditions: Combination of Multicast QoS configuration with BPI+ encryption coupled with the blocking of IGMP for those same multicast groups.
Workaround: Do not configure multicast QoS with BPI+ for groups where IGMP is blocked.
•
Symptoms: Multicast QoS service flow is not created for the DSG tunnel, after the CMTS reloads with the service class and DSG configuration.
Workaround: There is no workaround.
•
Symptoms: CPUHOG error messages appear when many multicast QOS sessions get deleted after IGMP leave within a short time.
Workaround: There is no workaround.
•
Symptoms: Cable line card crashes during CM registration in Remote Copy Protocol (RCP) verbose reporting mode.
Conditions: This issue is observed when the cable rcp-control verbose command is enabled and if the CM is sending the CLABS defined RCP 00 10 00 00 05 in the verbose format.
Workaround: Do not enable the cable rcp-control verbose command under the cable interface.
•
Symptoms: The show controller cable upstream command output for the MAC Parser queue information shows the same numerical value for the queue depth and wait time. The wait time is not shown correctly, the wait time is shown as a duplicate of the queue depth.
Workaround: There is no workaround.
•
Symptoms: The wideband SPA cannot upgrade firmware for a corrupted field-programmable gate array (FPGA) with field-programmable devices (FPD).
Workaround: If the SPA with the corrupt FPGA is plugged into the UBR10-2XDS-SIP, then the FPGA gets upgraded properly.
•
Symptoms: Narrowband cable modems cannot get online on remote downstreams after both the Protect and the Working line cards are reset in a CMTS with more than 2 SPAs in the system.
Workaround: Revert to the previous configuration or reload the CMTS.
•
Symptoms: The wideband SPA in Cisco SIP-600 is stuck after executing the hw-module bay x/y shutdown and no shutdown command.
Workaround: Reload the wideband SPA.
•
Symptoms: Executing the cmts walk command on ARP table causes high CPU utilization. This is also seen while using the snmp walk command on the ARP table.
Conditions: This issue occurs while using the cmts walk and snmp walk commands.
Workaround: To prevent high CPU utilization when using the snmp walk command, use snmp view commands:
–
–
–
–
•
Symptoms: Router is experiencing spurious memory access during the ipv6_spd_pak_classify.
Conditions: This issue is observed after configuring IPv6 on the Fast Ethernet interface.
Workaround: There is no workaround.
•
Symptoms: DSG:Downstream Channel Descriptor (DCD) counters increment even while modular interfaces are down.
Workaround: There is no workaround.
•
Symptoms: Memory leak is observed in cmts_fill_peak_offpeak_parameters.
Conditions: This issue is observed when the peak-offpeak enforce rule is changed to legacy settings and the rule is enabled.
Workaround: Do not change peak-offpeak rule to legacy and define a new enforce rule.
•
Symptoms: Traceback and spurious memory access is observed due to failover of either PREs.
Workaround: There is no workaround.
•
Symptoms: The SPA flaps after reloading the CMTS.
Conditions: This issue is observed on the Cisco router running Cisco IOS Release 12.2(33)SCB3 with 5 SFPs in slots 1 and 3 after reloading the CMTS.
Workaround: There is no workaround.
•
Symptoms: The ip dhcp relay info policy removal pad command is not ISSU compliant.
Workaround: ISSU cannot be conducted and the CMTS must be reloaded.
•
Symptoms: Though cable modems are online, the CMTS is not able to ping them after unrelated wideband or modular cable interface is shut down.
Workaround: Shut down the affected interface and then re-enable that interface.
•
Symptoms: CMTS crashes when removing the flash card.
Workaround: There is no workaround.
•
Symptoms: Traceback is observed after the Multicast client initiating igmp join with group-encryption is enabled.
Workaround: There is no workaround.
•
Symptoms: The interrupts are not disabled when a SPA is stopped.
Workaround: There is no workaround.
•
Symptoms: Multicast traffic was not sent when a RF channel is added to the existing multicast tunnel.
Workaround: Re-apply the DSG configuration or execute the shut and no shut command for the RF channel.
•
Symptoms: Ping failure observed after weekend traffic and the cable modem reset.
Workaround: There is no workaround.
•
Symptoms: Wideband modems go offline after line card switchover was triggered by cable power off.
Workaround: There is no workaround.
•
Symptoms: The IGMP input process causes high CPU utilization on the PRE.
Conditions: This issue is observed during high rate of IGMP joins received and multiple cable interfaces present on CMTS.
Workaround: Enable "multicast echo" on bundle interface.
•
Symptoms: After service flows are created for the multicast traffic, the show interface wideband 1/1/0:2 service-flow x counters command unexpectedly displays the service flow counters as zero.
Workaround: There is no workaround.
•
Symptoms: A few DSIDs are wrongly assigned to more than one wideband cable modem and ping failure occurs to one of them.
Workaround: Execute clear cable modem reset command to issue a new DSID.
Open Caveats—Cisco IOS Release 12.2(33)SCB3
•
Symptoms: Wideband modem downstream throughput rate is not yet supported.
Conditions: The packet per second and bits per second throughput rate for each wideband CM downstream service flow is not available. This affects both the show cable modem Cx/y/z service-flow [counters | verbose] and show cable modem a.b.c.d qos commands.
Workaround: There is no workaround.
•
Symptoms: String commands fail to run in the Cisco IOS Tcl shell.
Conditions: This issue occurs while running string commands in Cisco IOS Tcl shell.
Workaround: There is no workaround.
•
Symptoms: The throughput rate shows (slightly) more than the expected range.
Conditions: This issue happens when the Dynamic Message Integrity Check (DMIC) feature is enabled and the modem is marked with given maximum rate values. This can also happen when the QoS profile has got lower maximum rates.
Workaround: There is no workaround.
•
Symptom: The DOCSIS Set-Top Gateway (DSG) client does not work on the modular cable interface.
Conditions: This issue occurs on a fully functional and configured MAC domain on the Cisco uBR10-MC5X20 cable line card. When a modular cable SPA is inserted and configured, and the modular-cable downstream is applied to the MAC domain on the Cisco uBR10-MC5X20 line card, Downstream Channel Descriptor (DCD) messages are not sent.
Workaround: To enable the modular cable interface to function properly, it is necessary to execute the shut and the no shut commands on the MAC domain of the Cisco uBR10-MC5X20 line card that is associated with the modular cable interface.
•
Symptoms: Multicast stream drop, after changing the "ip igmp query-interval". The following show the output before and after the change:
Before the change:
Router# show ip mroute<<..skipped..>>(, ), 01:09:04/00:02:52, flags: sTIIncoming interface: TenGigabitEthernet1/1, RPF nbr 10.x.x.x, RPF-MFDOutgoing interface list:Vlan50, Forward/Sparse, 00:48:45/00:02:52, H <<====After the change:
Router# show ip mroute<<..skipped..>>(, ), 01:28:19/00:02:36, flags: sPTIncoming interface: TenGigabitEthernet1/1, RPF nbr 10.x.x.x, RPF-MFDOutgoing interface list: Null <<====Conditions: The issue is was first seen in a Cisco 7600 chassis with a Supervisor 720 engine running Cisco IOS Release 12.2(33)SRB.
Workaround: There is no workaround.
•
Symptoms: Inactive sessions are observed in the multicast explicit tracking database.
Conditions: This issue is seen when IGMP leave is received for a group followed by a CLC switchover.
Workaround: There is no workaround.
•
Symptoms: The cable line card crashes after the execution of the clear arp and clear cable modem delete commands.
Conditions: The condition is unknown.
Workaround: There is no workaround.
•
Symptoms: Certain PCMM calls using wideband CMs may have one way audio.
Conditions: This issue occurs when PCMM calls between w-online (pt) and online (pt) on wideband CMs have one-way audio.
Workaround: There is no workaround.
•
Symptoms: When Multicast Virtual Private Network (MVPN) is configured on the CMTS, and a CPE joins a multicast group that a source is already transmitting to, then the (S,G) entry is not created in the MVRF table of the CMTS.
Conditions: This issue occurs when MVPN is configured on the CMTS.
Workaround: There is no workaround.
•
Symptoms: The online cable modems go offline with no alert or log message.
Conditions: This issue is only related to Cisco uBR10-MC5X20U cards on the Cisco uBR10012 platform. This issue is similarly seen in Cisco uBR10-MC5X20H cards.
Workaround: Executing the shut or the no-shut commands on the affected interfaces or power-cycling the cable line card solves the problem.
•
Symptoms: When a modem comes up during off-peak time when it is not supposed to be monitored, as per the enforce-rule configured, the SNMP is not synced to the standby line card.
Conditions: This issue is seen if a line card switchover is done during off-peak time. Even if the modem enters the peak time, it is not monitored as per the enforced rule configured. The SNMP is synced to the standby line card once the modem is monitored during the peak time.
•
Symptoms: The CPE device disappears from the cable host table and sometimes from the ARP table, thus causing the CPE to lose IP connectivity. This shows up across different CM vendors, different CLC types, different plants, and on different Cisco uBR10012 routers.
Conditions: The issue is seen in the following:
–
–
–
–
Workaround: Execute clear cable host command for the CPE device until the device shows up on the CPE table again.
•
Symptoms: The upstream ports configured as HCCP protect show the port link down when executing the show facility-alarm status command.
Condition: This issue is seen when the cable line card is configured as HCCP protect, and the router is running a later image version than Cisco IOS Release 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: For deleted service flows associated with PCMM calls, the last character in the Service Class Name field is dropped in SAMIS records as well as in the SNMP MIB object "docsQosServiceFlowLogServiceClassName".
Conditions: This issue is seen when the dynamic service flows associated with PCMM calls are deleted. The last character is missing from the service class name in the MIB object "docsQosServiceFlowLogServiceClassName" and SAMIS records.
•
Symptoms: While executing the clear counter command, service flow statistics are not getting cleared under the wideband interface.
Conditions: This issue is seen in cases of multicast and the forwarding is through the wideband interface. Multicast service flows are created and the counters are maintained under the wideband interfaces.
Workaround: There is no workaround.
•
Symptoms: Memory leak at CMTS_HCCP_GET_TLV_PARSE_TLV has been observed at the RP.
Workaround: There is no workaround.
•
Symptoms: Both the working card and protect card are active.
Conditions: This issue is seen when a line card switchover happens from 5/0 to 5/1. After synchronization, shut down the 5/0/2 interface, and then revert from 5/1 to 5/0. If the 5/0/2 interface is not shut down, and after the suspend timer interval expires, both 5/0 and 5/1 are active.
Workaround: Do not shut down an active protect interface. If both the cards (working and protect) are active, reset the protect card.
•
Symptoms: Cable metering fails and enters a "hung" state.
Conditions: The issue happens when the "ip tcp timestamp" option is configured globally.
Workaround: Do not use the "ip tcp timestamp" option.
•
Symptoms: Traffic fluctuation only happens when the WB/MC topology and traffic rate are exactly the same.
Conditions: This issue is seen in Cisco IOS release 12.2SC with HQF scheduler. When 46 percent of bandwidth is allocated to Channel 0 for MC0 and 46 percent of the bandwidth is allocated to Channel 1 for MC1,the rest of the bandwidth is allocated to WB0.
Thus, the traffic rate of WB is higher than the configured percentage. The traffic rate of MC0 is a little lower than the configured percentage and the traffic rate of MC1 is a little higher than the configured percentage.
Workaround: Do the following:
–
–
–
•
Symptoms: During line card high availability, the CM instance bpi_enable flag on the Route Processor (RP) is not synced over correctly.
Conditions: This issue is seen while generating multicast stream to CMs, and if BPI CM is used during a line card switchover.
Workaround: There is no workaround.
•
Symptoms: The MAC Domain Descriptor (MDD) change message is generated on all the RF channels globally for any RF channel configured on the CMTS.
Conditions: This issue occurs when an MDD change is supposed to be generated for channels within the MAC domain.
Workaround: There is no workaround.
•
Symptoms: When a Cisco uBR10012 router has a lot of CMs, the show cable modem docsis version summary command total may not be able to display each column properly because there is not enough space between columns.
Conditions: This issue is seen when there are a lot of cable modems. This issue is seen in Cisco IOS Release 12.3(23)BC and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: When the cable sync-interval command is configured, the cable interface needs to execute a shut or no-shut command for the interval change to take effect.
Conditions: This issue occurs while configuring the cable sync-interval command.
Workaround: There is no workaround.
•
Symptoms: The downstream Committed Information Rate (CIR) reservation for wideband cable modems works as expected, however they are displayed incorrectly under admission control command.
Conditions: This issue is seen in any w-online state wideband cable modem with non-zero CIR reservations. It displays the CIR reservation as 0 although the reservation admit or not admit works properly. The show interface admission-control reservation downstream command displays the 0 CIR for the wideband CM.
Workaround: Use the show cable modem qos command to display the correct reservation.
•
Symptoms: When either Cisco uBR-MC5X20U/S/H card MAC domains have wideband associations, a traceback may be observed along with a message similar to the following when NO CARD x/x is used:
*Jan 25 05:41:57.235 PST: %SYS-3-BADLIST_DESTROY: Removed a non-empty list(2F3F7BF4, name: (unknown)), having 36 elements -Process= "C10K Card Actions", ipl= 0, pid= 75A small amount of memory leak may also happen. No other side effects are expected.
Conditions: This traceback is seen only if the card had a NB-WB association previously, and was powered down or removed.
Workaround: There are no workarounds.
•
Symptoms: After a PRE processor module switchover, the following message is displayed:
"*Jan 29 10:22:33.187: %IPC-5-INVALID: Message Port index Dest Port 0x1001C Session 0x1 Source 0x60000"Conditions: This issue has been observed on In Service Software Upgrade (ISSU) while upgrading PRE A slot to PRE B slot processor modules.
Workaround: There is no workaround.
•
Symptoms: Port-channel interfaces used in conjunction with 802.1q subinterfaces show a number of problems on both the PRE4 and PRE2 processor modules. This was observed while testing the Cisco SPA-5X1GE-V2 in Cisco IOS Release 12.2(33)SCB. The problem appears to be with the 4-jacket WAN interface, and not with the older half-height Gigabit Ethernet interfaces.
Conditions: This issue occurs in a PRE2 processor module under the following conditions:
–
–
This issue occurs in a PRE4 processor module under the following conditions:
–
–
Workarounds: To restore the traffic, do the following:
1.
2.
3.
4.
Theoretically, you can boot the CMTS with no Ethernet channel configured, and copy an edited production configuration file to running-config after all the interfaces are up.
•
Symptoms: The Cisco Broadband Troubleshooter trace window appears to show incorrect data when triggering an upstream by CM MAC address. The MIB object "server% getmany -v2c ccsSpectrumDataPower" returns values that when graphed do not show the expected QAM haystack. This MIB does not provide data until the CMTS is first configured appropriately via SNMP sets.
Conditions: This issue occurs in upstream frequency stacking and configured MAC domains.
Workaround: Remove the frequency stacking and configured MAC domains.
•
Symptoms: The w-online state modem comes online after protect line cards are reset before a resynchronization.The CMTS is configured with N+1 redundancy. If the working line card (7/0) is reset, the modem switches to protect (5/0). If the protect (5/0) line card is reset before resynchronization, the modem disappears, and comes in online instead of being in the w-online state.The clear cable modem reset command does not work to bring the modem back to the w-online state.
Conditions: This issue occurs while resetting an active line card during the standby line card bootup period, but before the static synchronization of an active to standby line card finishes.
Workaround: Reload the line card or the CMTS.
•
Symptoms: An IRONBUS_FAULT is displayed while pulling out the Cisco SIP-600.
Conditions: This issue is seen in Cisco IOS Release 12.2(33)SCB image running on a Cisco uBR10012 router.
Workaround: There is no workaround.
•
Symptoms: A Spurious Memory access was observed on a cable line card when Online Insertion and Removal (OIR) of the Cisco SIP-600 was done on a PRE 4 processor module.
Conditions: This issue is observed on PRE 4 processor module with an active wideband SPA.
Workaround: There is no workaround.
•
Symptoms: The show cable bundle multicast command does not display the entire bundle in the interface column.
Conditions: This issue is seen in the output of the show cable bundle multicast command.
•
Symptoms: Multimedia Terminal Adaptor (MTA) CMs are not accessible from the CMTS or Internet because the ARP entry is replaced with an all-zero MAC address. Hence it is not accessible when a crafted packet is sent.
Conditions: This issue occurs when a crafted packet is sent to the CMTS, causing the CMTS to use the wrong MAC address for the ARP table.
Workaround: Execute the cable source-verify dhcp server x.x.x.x command, where x.x.x.x is your DHCP server, and use cable arp filter to limit the potential ARP storm.
•
Symptoms: Wideband CMs using a stats index are not managed by the modular-host. This causes two modems using the same Blaze header stats index, thus leading to incorrect downstream traffic statistics on these modems.
Conditions: This issue occurs when there are wideband modems in the w-online state and a change is made to the downstream channel ID of the MAC domain host of these wideband modems.
Workaround: There is no workaround.
•
Symptoms: The output of "sid" is incorrect in the show cable modem x.x.x.x service-flow [verbose] command.
Conditions: This issue is seen in Cisco IOS Release 12.3(23)BC and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptom: The title field "Name" in the show cable admission-control interface is missing.
Conditions: This is seen in the output of the show cable admission-control interface command.
Workaround: There is no workaround.
•
Symptoms: Traceback and IPC error messages may appear after the CMTS is loaded with the Cisco IOS Release 12.2(33)SCB1 image.
Conditions: This issue is seen in routers running Cisco IOS Release 12.2(33)SCB1.
Workaround: There is no workaround.
•
Symptoms: The show pxf cable multicast-echo does not display the entire interface when the subinterface is greater than 13 characters.
Conditions: This issue is seen for subinterfaces greater than 13 characters.
•
Symptoms: Modifying the second-choice-channel-width has no affect on the configuration even though the cable upstream channel-width command accepts the input.
Conditions: This issue is seen in Cisco IOS Release 12.3BC.
Workaround: Modify the first-choice-channel-width, and then reconfigure the first and second choice.
•
Symptoms: The packet drop counter is not getting updated in show interface service-flow counter for single-session multicast service-flows
Conditions: This issue is seen in single session multicast service-flows.
•
Symptoms: No multicast-session is permitted when cable multicast authorization profile (mauth) is enabled and the debug messages display "session limit reached".
Conditions: This issue is seen in the following conditions:
–
–
–
–
–
Workaround: Delete the cable modem and allow the modem to come online with second cable modem configuration (non-mauth cm-file).
•
Symptoms: The wideband cable interface is not able to pass traffic.
Conditions: The issue occurs because the bonded channel-to-RF channel mapping configuration is missing from the SPA.
Workaround: There is no known workaround.
•
Symptoms: The CMTS may initiate DCC/DSA/DSC when Dynamic Bonding Change (DBC) is outstanding.
Conditions: If the customer is running Video over DOCSIS (VDOC) and Cisco Load Balancing (or CMTS initiated DSA) with a DPC3000, the CMTS makes VDOC higher priority over load balancing (or CMTS-initiated DSA) when a DBC is outstanding. According to DOCSIS 3.0 specifications, the CMTS must not issue a DCC or DSA/DSC command when it has previously issued a DBC command and that command is still outstanding.
Workaround: There is no workaround.
•
Symptoms: CMTS does not reject the CM-initialized DSC/DSA messages during Dynamic Bonding Change (DBC) or Dynamic Channel Change (DCC).
Conditions: This issue is seen if the customer is running Video over DOCSIS (VDOC) with a Cisco DPC3000 cable modem, and the CM is making a DSA call. The CMTS responds with a confirmation code of "reject-temporary" for the DSA-REQ or DSC-REQ for Annex B.
Workaround: There is no workaround.
•
Symptoms: After issuing the no cable trust command on the active PRE module, that same command is not removed from the standby PRE module and as a result the active and standby PRE modules are out of sync.
Conditions: This behavior is only seen if the MAC address used in the cable trust command does not exist in the CMTS.
Workaround: There is no workaround.
•
Symptom: Traffic does not resume after a PRE module switchover and modems are not accessible after a PRE module switchover.
Conditions: This issue occurs after a PRE module switchover and "cef"is not running in the standby PRE module.
Workaround: There is no workaround.
•
Symptoms: After unconfiguring the cable source-verify dhcp command, the "SrcVfy" field continues to display "On".
Conditions: This is seen when the cable ipv6 source-verify and cable source-verify dhcp commands are configured on the same bundle interface, and when the cable source-verify dhcp command is unconfigured before the cable ipv6 source-verify command.
•
Symptoms: Upstream-and-downstream-related tables are not retrieved while querying via SNMP.
Conditions: This issue is seen while querying via SNMP.
Workaround: There is no workaround.
•
Symptoms: When configuring cable upstream threshold cnr-profile snr-profile, the parser shows and accepts the second threshold value. It does not give an error message, and does not display the configuration with only the first threshold value.
Conditions: This issue is seen while configuring the CNR and SNR thresholds for modulation profiles.
Workaround: Either enter any valid threshold value or "0" to bypass.
•
Symptoms: The value range of the MIB object "ccmtrCollectionInterval" is not consistent with the command. The valid range for MIB object is 15 to 1440, but for the command, the range is 2 to 1440.
Conditions: This issue occurs while setting the MIB object "ccmtrCollectionInterval" using the SNMP set operation.
Workaround: Set the interval using the command.
•
Symptoms: A memory leak was observed after deleting the subinterface.
Conditions: This issue is seen while deleting a subinterface.
Workaround: There is no workaround.
•
Symptoms: When applying modulation profiles to an upstream port, the modulation profiles are re-ordered, high performance to robust/low performance. Modifying the order of the modulation/symbol rate on the profile does not happen.
Conditions: If the signal-to-noise ratio (SNR) threshold is set at 25 with a hysteresis of 2 (bypassing all FEC), the modulation on the upstream will be QPSK until the SNR falls below 25 dB where it will change to 16-QAM. When the SNR reaches 27 dB (25 dB + 2 dB), the modulation reverts to QPSK. When a line card switchover is performed, the modulation profiles are re-ordered on the protect card.
Workaround: Remove and re-apply modulation-profile configuration on the upstream port after modifying the modulation profiles.
•
Symptoms: All CMs on adjacent downstreams are offline. The adjacent downstream ports are 0 and 1, or 2 and 3, or 4.
Conditions: The downstream ports that share JIB may be nonfunctional if the JIB that services both downstream ports is locked.
Workaround: Reload the cable line card.
•
Symptoms: A memory leak was observed after executing the multicast test case.
Conditions: This issue is seen when the cable multicast group-qos and cable multicast qos group priority commands are unconfigured.
Workaround: First unconfigure the session-range using the cable multicast qos group priority command.
•
Symptom: Removing the line card configuration causes %LINEPROTO-5-UPDOWN and scm sum total, scm vendor sum, and scm show modem commands show no CMs.
Conditions: This is seen when line card high availability is removed and while executing scm sum total, scm vendor sum. and scm show modem commands.
Workaround: There is no workaround.
•
Symptoms: PXF forwards traffic up to 300 Mbps to the Route Processor (RP) using 1500-byte IPv6 packets.
Condition: This issue occurs when Physical Logical Bandwidth Limiting Traffic (PBLT) for the RP CPU queue is set to 300 Mbps. The show pxf cpu queue command displays this information.
Workaround: Change the value to 44 Mbps (DS3 Bandwidth) for the PRE2 processor module and 1.6 times 44 Mbps for the PRE4 processor module.
•
Symptoms: The MIB object "entPhysicalName" for the Cisco uBR10012 router does not display the values correctly. The values as displayed as slot 0/0/0 and slot 0/1/0 for the PRE processor modules instead of being displayed as 0A and 0B from left to right. Similarly, the MIB object "entPhysicalName" returns slot x/x/x (for example, 1/0/0) instead of the values displayed in the device.
Conditions: This issue occurs while querying the MIB object "entPhysicalName".
Workaround: There is no workaround.
•
Symptoms: When applying new RF channels in Cisco IOS Release 12.2(33)SCB, a corresponding modular-cable interface is created. These interfaces appear under router RIP, router OSPF, and router isis routing protocol processes as "no passive-interface", even when passive-interface default is configured. Cable, modular-cable, and wideband interfaces should not appear under a routing protocol process as there is no Layer 3 control.
Condition: The creation of new cable bundle interfaces is configured rightly using the passive-interface default command. Because there is no Layer 3 control on cable, modular-cable, or wideband cable interfaces, such a problem with the Layer 2 interfaces appearing as "no passive-interface" does not lead to improper routing protocol control, but this can create some confusion even though the issue is cosmetic.
•
Symptoms: A crash is observed on the CMTS.
Conditions: This issue is seen in cases ETDB show commands are paused and multicast sessions removed because a CPE leave or a CM goes offline.
Workaround: Set the term length to 0 for no pausing while displaying the command.
•
Symptoms: The CMTS allows any upstream to be attached to any connector without any warning message. A "no" is displayed for the conflicting connector.
Conditions: This issue is seen when a new card is plugged into a slot with no pre-configuration.
Workaround: Reload the line card or use default connectors.
•
Symptoms: The range for Security Association Identifier (SAID) is not correct. The correct range should be 8192 to 12272.
Conditions: This issue is seen in routers running Cisco IOS Release 12.2(33)SCA.
•
Symptoms: Different values for the MAC Domain Downstream Service Group (MD-DS-SG) are seen after an RPR switchover, using the show cable modem wideband and show cable mac-domain downstream-service-group commands.
Conditions: This issue occurs while changing the MD-DS-SG by changing the fiber node when CMTS is up and after a RPR switchover.
Workaround: There is no workaround. This symptom is not critical. It does not affect normal use of relative features.
•
Symptoms: The PRE4 processor module rate shaping allows the traffic throughput on a service flow to exceed its maximum rate.
Conditions: This issue is seen on PRE4 processor module best effort service flows.
Workaround: There is no workaround.
•
Symptoms: A crash is observed on the ESR-PRE2 processor module. The crash information logs may contain the following:
%UBR10K-6-US_SFID_INCONSISTENCY: US-SF found: SFID xxxx, type 0, sid 0(yyyy), MAC aaaa.aaaa.aaaa(bbbb.bbbb.bbbb), prim_sid xxx(yyy)CMD: 'no cable service attribute voice-enabled downstream-type HA-capable'Conditions: This issue is seen in Cisco IOS Release 12.3(23)BC4 and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: The show controller modular-cable command output displays invalid voltage measurement readings for the 24 rf-channel SPA.
Conditions: This is seen on the 24 rf-channel SPA.
Workaround: Re-execute the show controller modular-cable command.
•
Symptoms: The cable monitor duplicates one more MAC Domain Descriptor (MDD) for every generated MDD if it is configured to monitor cm_multicast_addr(01e0.2f00.0001).
Conditions: This issue is seen when cm_multicaset_addr(01e0.2f00.0001) is configured.
Workaround: Configure the paktype mac command to capture MAP/MDD/UCD.
•
Symptoms: After a reload, cable metering source-interface FastEthernet0/0/0 is not present at the running configuration, although cable metering filesystem disk0: cpe-list-suppress and cable metering source-interface FastEthernet0/0/0 were configured.
Conditions: This issue is seen only in Cisco IOS release versions on the PRE2 processor module and not in Cisco IOS releases versions of the Cisco uBR7200 router.
Workaround: There is no workaround.
•
Symptoms: Delays occur in calls after a failover from a Cisco uBR10-MC5X20S card to a Cisco uBR10-MC5x20H card.
Conditions: This issue is seen in a Cisco uBR10012 router running Cisco IOS Release 12.2(33)SCB1 and configured for N+1 Redundancy and after a failover of a Cisco uBR10-MC5X20S card to Cisco uBR10-MC5X20H card. A call take approximately 20 seconds to successfully hang up. Before that period, if the handset is placed on the hook and then picked up, dial tone would not be heard.
Workaround: There is no workaround.
•
Symptoms: The active DOCSIS queues remain active in the event of a DOCSIS SPA failure.
Conditions: If there are modems online and passing traffic on the modular-cable and/or on wideband-cable interfaces and the DOCSIS SPA fails, the DOCSIS queues associated with these online modems remain active attempting to dequeue.
Workaround: There is no workaround.
•
Symptom: When attempting to upgrade a Cisco uBR10012 router from Cisco IOS Release 12.3(21)BC to Cisco IOS Release 12.2(33)SCB, the primary PRE may crash.
Conditions: This issue is seen when the CMTS is running Cisco IOS Release 12.3(21)BC and while upgrading to Cisco IOS Release 12.2(33)SCB. The CMTS must then be powered down and physically reconfigured to run as a M-CMTS (Cisco SIP-600 inserted, WB SPAs inserted, DTCCs inserted, so on.). Lastly, the CMTS must be turned on to run its new Cisco IOS Release 12.2(33)SCB IOS image and physical configuration.
•
Symptoms: The RFGW-1 reports channel(s) as down following rf-channel and fiber node configuration changes.
Conditions: This issue is seen when there are changes made to the rf-channel/fiber node configuration on the RFGW-1. It reports the affected channels as DOWN.
Workaround: Exit fiber node configuration level and the changes are be committed and the rf-channel status as reported by the RFGW-1. It shows the correct state.
•
Symptoms: When attempting to upgrade a Cisco uBR10012 router from Cisco IOS Release 12.3(21)BC to Cisco IOS Release 12.2(33)SCB, the secondary PRE processor module may crash directly after the primary PRE processor module crashes.
Conditions: This issue occurs in routers running Cisco IOS Release 12.3(21)BC.
•
Symptoms: Modems marked by "tftp-enforce mark-only" show up in show cable modem as marked with the # sign. Syslog messages might report that the modem was marked as follows:
SLOT 6/1: Mar 25 12:21:58.967 CDT: %UBR10000-4-REGISTRATION_BEFORE_TFTP_MARK: Registration request unexpected: Cable Modem did not attempt TFTP. Modem marked with #. CM Mac AddrBut the modem does not show up with the "#" in front of the online status in show cable modem command.
Conditions: This is seen in routers running Cisco IOS Release 12.2(33)SCB1.
Workaround: There is no workaround.
•
Symptoms: The show facility-alarm status command displays a false alarm.
Conditions: This issue is seen for the protect line cards when connectors are not set.
Workaround: There is no workaround.
•
Symptoms: When a bundle subinterface number is greater than 64K, the statindex assignment for multicast session shows 0.
Conditions: This is a rare situation and only happens when a subinterface number is greater than 64K.
Workaround: Use a smaller subinterface number.
•
Symptoms: The CMTS sends DBC to the CM to change the primary downstream port
Conditions: This issue is seen when the VDOC broadcast feature is configured such that multiple RCCs in the same MAC domain have exact same frequencies for the primary bonding group, but the primary frequency is different. For example, if you have frequencies f1 and f2 as the primary bonding group frequencies, then rcc1 (f1f2f3) and rcc2 (f1f2f4) have the first two frequencies as f1f2, while rcc3 (f2f1f3) and rcc4 (f2f1f4) have the first two frequencies as f2f1. With this configuration, the CM is first online using rcc3 and the CPE sends an IGMP join for the channel corresponding to f3.
Workaround: For the VDOC broadcast feature, configure RCC templates such that a set of primary bonding group frequencies are in the same order in all RCC templates.
•
Symptoms: Facility alarms are incorrectly generated for cable line card interfaces.
Conditions: This issue occurs when line card redundancy is configured or upstream frequency stacking is configured.
Workaround: There is no workaround.
•
Symptoms: Executing the cmts walk command on ARP table causes high CPU utilization. This is also seen while using the snmp walk command on the ARP table.
Conditions: This issue occurs while using the cmts walk and snmp walk commands.
Workaround: To prevent high CPU utilization when using the snmp walk command, use snmp view commands:
snmp-server view cutdown iso included
snmp-server view cutdown at excluded
snmp-server community public view cutdown ro
snmp-server community private view cutdown rw
•
Symptoms: DMIC code is being corrupted due to a DHCP packet.
Conditions: This issue seen on PRE2 processor module running Cisco IOS Release 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: Although the CM is in non-BPI mode, after line card high availability, its downstream packet or the ping are still encrypted and fail.
Conditions: This issue is seen in the following conditions:
–
–
–
Workaround: After the change to non-BPI configuration, use the delete command instead of reset.
•
Symptom: The CM goes offline after line card fails with different types of working and protect line card (for example, Cisco uBR10-MC5X20H protect and Cisco uBR10-MC5X20U line cards)
Conditions: This issue occurs when a manual line card switchover is done and when a new active line card crashes or reboots before HCCP Static Sync is done.
Workaround: There is no workaround.
•
Symptoms: The multicast traffic is not forwarded by the CMTS to the multicast clients after the IGMP join request is sent by the client. The "packets received" counter keeps increasing, but the "packets forwarded" counter stops increasing
Conditions: The issue is seen with Cisco uBR10-MC5X20H line cards with multicast configuration enabled.
Workaround: Configure the ip igmp version 3 command in the bundle interface associated with the multicast traffic forwarding line card.
•
Symptoms: When the upstream service flow of modem exceeds the 1 MB limit on a Cisco DPC3000 cable modem, no traffic can be sent.
Conditions: This happens when the upstream traffic is sent at 1 MB or above.
Workaround: There is no workaround.
•
Symptoms: After a line card switchover the key-index becomes 0 for a multicast QoS that is configured for a active static IGMP session along with encryption.
Conditions: This issue is seen only for static IGMP groups which has multicast QoS with encryption configured.
Workaround: There is no workaround.
•
Symptoms: An unused Downstream Service Identifier (DSID) and statsindex is shown in the show cable multicast dsid command. The Service Flow Identifier (SFID) value displays "0" in the show interface multicast-session command.
Conditions: This happens when admission control checks fails at the cable line card (CLC) during multicast service flow creation.
Workaround: There is no known workaround.
•
Symptoms: If a modular-cable channel with online modems is removed from its host and reconfigured to another host, the flows counter are incorrectly displayed.
Conditions: This is seen when a modular-cable channel is removed from its host and reconfigured to another host.
Workaround: There is no workaround.
•
Symptoms: The Gigabit Ethernet interface on an ESR-1GE may fail to send packets.
Conditions: This might be seen just after executing online insertion and removal (OIR) of an ESR-1GE with low rate.
Workaround: Execute the hw-module slot reset command.
•
Symptoms: The multicast packets are forwarded to a cable downstream interface that is in rf-shutdown state.
Conditions: This issue occurs if the cable interface is in rf-shutdown state and multicast traffic is forwarded to that interface unnecessarily.
Workaround: There is no workaround.
•
Symptoms: The fiber node accepts non-existent upstream connectors, which may confuse the user on physical and logical topology.
Conditions: This issue occurs when configuring non-existent upstream connectors on the fiber node.
Workaround: Do not configure non-existent upstream connectors for the fiber node.
•
Symptom: The CMTS causes high CPU utilization on the IPDR/SAMIS collector by sending multiple records for a flow with the same creation time.
Conditions: The service flow export and service flow deletion occur at the same time.
Workaround: There is no workaround.
•
Symptoms: Multicast traffic forwarding stops on modular cable interfaces.
Conditions: This issue occurs on modular downstream interfaces with multicast QoS functionality configured and the Parallel Express Forwarding (PXF) on the CMTS has restarted.
Workaround: In case of a static multicast, unconfigure and reconfigure the ip igmp static-group commands. In case of dynamic multicast, use the clear ip mroute command.
Workaround: There is no workaround.
•
Symptom: Logged messages are sent not only to the IOS system log in memory on the Cisco MC5X20 line cards, but to the console serial port. This I/O activity uses CPU resources unnecessarily especially at card bootup time. Console logging may be re-enabled by configuring the logging console command.
Conditions: This issue is seen in all Cisco uBR10-MC5X20 line cards.
•
Symptoms: The Cisco uBR10012 router crashed because it ran out of un-fragmented memory.
Conditions: This is seen in Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptom: The DOCSIS Timing, Communication and Control (DTCC) card information is displayed in the output of the show diagnostic command when the card is not present.
Conditions: This issue is seen when both SPA cards are present in the slot/subslot and utility cards are present in slot/subslot. For example, a SPA is present in 1/1 and a DTCC card is present in 1/1.
Workaround: There is no workaround.
•
Symptoms: Stale multicast BPI encryption keys remain on protect line cards that are in standby state.
Conditions: This issue is seen while joining and leaving a multicast group (present in cable match command). Conduct a line card switchover from working to protect and then revert back to working from protect.
Workaround: There is no workaround.
•
Symptoms: A multicast QoS session is not automatically created when changing group configuration or GQC.
Condition: This issue is seen when a multicast QoS group is configured on a bundle interface and a matching static-group is defined. However, any change in the group configuration after it is applied to the bundle interface is not taken into effect, and the original multicast QoS session is removed.
Workaround: Remove the multicast QoS session from the bundle interface and re-apply it to the bundle.
•
Symptoms: The show controller modular-cable 1/0/0 command displays that rf-channel 0 becomes unknown after a SPA hw-module reset.
Conditions: This issue occurred executing the hw-module reset and show diagnostic 1/0/0 commands. This causes all wideband modems also to go online.
Workaround: Reload the router to fix the issue.
•
Symptom: The Multimedia Terminal Adapter (MTA) lacks IP connectivity and cannot be reached.
Conditions: This is seen when upgrading to Cisco IOS Release 12.2(33)SCB2.
Workaround: Reset the MTA.
•
Symptom: The "PXF DMA Cumulative Ironbus Error Statistics" values are incriminated when executing the show pxf dma counters | include Ironbus command. For example, the error statistics displayed are:
PXF DMA Cumulative Ironbus Error Statistics:Ironbus Event 1/0: 0Ironbus Event 1/1: 166Ironbus Event 2/0: 2843674Ironbus Event 2/1: 2843674Ironbus Event 3/0: 70Ironbus Event 3/1: 67Ironbus Event 4/0: 63Ironbus Event 4/1: 64Ironbus Event 5/0: 241Ironbus Event 5/1: 2843674Ironbus Event 6/0: 278Ironbus Event 6/1: 268Ironbus Event 7/0: 250Ironbus Event 7/1: 269Ironbus Event 8/0: 2843674Ironbus Event 8/1: 1580072Conditions: The above errors are observed on a Cisco uBR10012 router with a PRE2 processor module running Cisco IOS Release 12.2(33)SCB1 and Cisco IOS Release 12.2(33)SCB2.
Workaround: There is no workaround.
•
Symptoms: The Arris Embedded Multimedia Terminal Adapter (EMTA) modems do not get a dial tone after a PXF crash.
Conditions: This issue is seen on a Cisco uBR10012 router running Cisco IOS Release 12.2(33)SCB2. This is seen after a PXF crash and when the Arris modems are in the "In Progress Telephony-DHCP" state.
Workaround: Switch over to the standby PRE module and execute the clear cable modem command.
•
Symptoms: A line card switchover from a Cisco uBR10-MC5X20U card to a Cisco uBR10-MC5X20H card causes only Cisco DPC3000 modems to go offline.
Conditions: This issue is seen in Cisco DPC3000 cable modems.
•
Symptoms: During the VDOC video channel changes, it is possible that the multicast QoS service flow cannot be created because a stale multicast QoS service flow exists for 140 seconds after the last IGMP member. This prevents a new multicast QoS service flow from being created.
Conditions: This is seen in multicast QoS service flows.
•
Symptoms: The "Gate Report State" counter in the output of the show packetcable cms verbose command does not increment for PacketCable Multimedia (PCMM) policy servers.
Conditions: This issue is seen for PCMM policy servers.
Workaround: There is no workaround.
•
Symptoms: The PacketCable Multimedia (PCMM) calls fail as the Gate Set Ack/Err is not received.
Conditions: This issue is seen in PCMM calls with a small value used for timer T1 (such as 1 second).
Workaround: Use larger values for the PCMM timer T1 (at least 5 seconds).
•
Symptoms: A a long latency for VDOC video channel change is seen as the Dynamic Bonding Change (DBC) timer restricts the CMTS to have another transaction on the same CM.
Conditions: This is seen due to the DBC timer value.
Workaround: Lower the T10 timer to 10 ms.
•
Symptom: CMTS will accept inconsistent "per-SF L2VPN TLV" in the cable modem configuration file.
Conditions: This is seen in the following conditions:
–
–
–
Workaround: There is no workaround.
•
Symptom: The Policy Based Routing (PBR) functionality (set ip next-hop and set ip default next-hop) does not work with the PRE4 processor module.
Conditions: This is seen in the PRE4 processor module.
Workaround: There is no workaround.
Resolved Caveats — Cisco IOS Release 12.2(33)SCB3
•
Symptoms: While viewing a modem upstream trace with Cisco Broadband Troubleshooter Release 3.2, the burst is not be displayed when equalization is configured on the US port. This seems to be happening with the Cisco uBR10-MC5X20S and Cisco uBR10-MC5X20U cards only.
Conditions: This issue is only on the Cisco uBR10-MC5X20S and Cisco uBR10-MC5X20U cards and when PRE-EQ (equalization-coefficient) is configured.
Workaround: Do not configure the PRE-EQ feature. This feature is off by default.
•
Symptoms: A PRE module failover can result in tracebacks and errors when the DOCSIS trust points are configured.
Conditions: Simply attempting a failover will result in failure.
Workaround: Remove the trustpoint keys for the DOCSIS-US-TRUSTPOINT and DOCSIS-EU-TRUSTPOINT. This keeps the trustpoints in place but does not associate any crypto keys with it.
•
Symptoms: The Cisco uBR10-MC5X20H card reloaded during forced line card failover.
Conditions: This is seen in the Cisco uBR10-MC5X20H.
Workaround: There is no workaround.
•
Symptoms: When a router (with a large-scaled configuration) is provisioned to perform a PRE module failover and a PXF crash occurs (for example, issuing a switchover pxf restart 1 0 command), this could result in missed IPC keepalives and a large crash information file being written, causing unexpected behavior including line card reloads.
Conditions: This issue is seen in a router that has a large configuration and has redundancy, mode switchover, main-cpu and switchover pxf restart 1 0 commands configured.
•
Symptoms: The ip tcp adjust-mss command results in packet loss for non-TCP traffic.
Conditions: This issue is seen when using the ip tcp adjust-mss command.
Workaround: Disable ip tcp adjust-mss on all interfaces in the device. This may cause higher CPU due to fragmentation and reassembly in certain tunnel environments where the command is intended to be used.
•
Symptoms: The Cisco uBR10012 router does not properly re-initialize the sequence number field in the downstream extended header of the first packet transmitted on a newly created re-sequencing Downstream Service Identifier (DSID). The CMTS must use a Packet Sequence Number of zero (0) in the DS EHDR of the first packet transmitted on a newly created re-sequencing DSID.
Conditions: This is seen in wideband when multiple downstream primary and secondary flows are used.
Workaround: There is no workaround.
•
Symptoms: An SNMP v3 walk with authentication does not complete against a virtual switch system (VSS).
Conditions: This issue occurs on an SNMP v3 walk with authentication on a VSS.
•
Symptoms: Modems on a modular interface suffer long-time recovery time or even fall offline after switchover or revertback.
Conditions: This only affects Cisco uBR-MC5X20S/U card as the active card. It affects modems on the modular interface and is more visible for Annex A modulation type.
Workaround: There is no workaround.
•
Symptoms: On the Cisco uBR10012 platform, during PRE module runversion of the ISSU upgrade process, the IPC connection between the RP and the cable line cards may take additional seconds to come up.
Conditions: This issue happens during PRE module runversion.
Workaround: There is no workaround.
•
Symptoms: A warning message such as the one below is erroneously printed during ISSU, even when actual CPU utilization is not high:
%HCCP_LC_ISSU_SHIM-3-RP_HIGH_CPU_UTILIZATION: High CPU utilization on PRE, line card ISSU may fail.Conditions: This happens during ISSU load version.
•
Symptoms: A crash occurs due to a bus error after displaying the following error messages:
%DATACORRUPTION-1-DATAINCONSISTENCY: copy error,%ALIGN-1-FATAL: Illegal access to a low address < isdn function decoded>Conditions: This crash first was seen on a Cisco 3825 running Cisco IOS Release12.4(22)T with ISDN connections going on. At the moment of the crash, the customer was monitoring the router using an SNMP Get program. After some minutes, the router reloaded by itself.
Workaround: There is no workaround.
•
Symptoms: A memory leak at buffer_peak_notify is observed.
Workaround: There are no workarounds.
•
Symptom: When Three Way Calling (TWC) is used, the admitted service flow of the on-hold side still has traffic going through.
Conditions: This is seen in three-way calling when one side is on hold. This is seen in Cisco IOS Release 12.3(23)BC and Cisco IOS Release 12.2(33)SCB release.
Workaround: There is no workaround.
•
Symptoms: The SP messages can get logged on the RP console but are not sent over as SNMP traps by the RP SNMP agent.
Conditions: This was first seen in Cisco IOS Release 12.2(18)SXF.
Workaround: There is no workaround.
•
Symptoms: Cable metering collection enters the "write-error" state and does not recover.
Conditions: This issue occurs when cable metering is configured with default TCP parameters.
Workaround: Enabling "ip tcp path-mtu-discovery" may help prevent occurrences of the issue. Also, the test cable metering abort command will clear the "hung" state and allows the next iteration of cable metering to occur.
•
Symptoms: The show interface cable x/y/z sid command shows null MAC address and null IP address SID entries. The show cable modem summary total command reports many more offline modems than are actually offline.
Conditions: This is observed in Cisco IOS Release 12.2(33)SCB.
Workaround: Clear the offline modems using the clear cable modem offline delete command. This issue is cosmetic.
•
Symptoms: A cable modem may disappear after a line card switchover.
Conditions: This issue occurs when configuring the member subslob x/y working command and when this configuration already exists and the working card x/y has downstream modular a/b/c RF configured.
Workaround: Execute a shutdown command followed by a noshutdown command on the corresponding modular interface.
•
Symptoms: Erroneous routing prefixes may be added to the routing table.
Conditions: This is seen when the DHCPv6 relay feature is enabled and a router receives a normal DHCPv6 relay reply packet, and this may lead to an erroneous route being added to the routing table.
Workaround: There is no workaround.
•
Symptoms: When issuing the show cable multicast db command with a page break, there may be a crash when there are IGMP activities undergoing.
Workaround: Use the term len 0 to print all data at one time to prevent this race condition.
•
Symptom: The following warning is seen when wideband (WB) resiliency is enabled, and the modem is moved from WB to NB.
%GENERAL-2-CRITEVENT: Request CIR exceeds available link rate.%UBR10K-3-QALLOCFAIL: Failure to allocate QoS queue for service flow 34, CM 001e.6bfa.f67aConditions: This issue is seen when wideband (WB) resiliency is enabled. The primary WB channel fails to work, and the modem is removed from WB to NB to recover service. The modem is configured with non-zero CIR. The NB reserved bandwidth is not enough to add a new modem.
Workaround: Disable WB resiliency.
•
Symptoms: After clearing all modems on c5/0/0, statistics on SPA downstream service flows may not work correctly.
Conditions: This issue is seen while executing the clear cable modem command causing the SPA statistics information on the SIPs to be removed.
•
Symptoms: The IPDR can trigger high RP CPU utilization when entries in the service flow log table time-out, and the data is written to the disk.
Conditions: This issue is seen when IPDR uses the file system for the deleted service flow information if the cable sflog command is configured such that the entry-duration is too short or the number of entries is too low.
Workaround: Prevent storage of deleted service flows on the disk by using the recommended configuration of the cable sflog command. The entry duration should be at least four times the metering interval and the max-entry should be large enough to store the deleted flows that exist in a metering interval.
cable metering destination < retries> 15 non-securecable sflog max-entry 30000 entry-duration 3600With the metering interval at 15 minutes, the service flow log should be at least 1 hour or 3600 seconds. The high CPU occurs when the file system is full, and attempts to write are continued.
•
Symptoms: Multiple memory leaks are seen at Cisco 10000 router On Board Failure Logging (OBFL) process.
Conditions: This is seen in the Cisco 10000 router OBFL process.
•
Symptom: The CMTS metering collection process on the line card or Route Processor (RP) gets terminated after the metering is completed. If destination-based CMTS-metering is enabled, and the export fails, a traceback is received around the data collection intervals. No tracebacks are reported if the export succeeds.
Condition: This could be an issue in the production environment. If the destination collector stops responding on the port or if it is unavailable, then a traceback is received on the CMTS.
•
Symptoms: When a PRE processor module switchover occurs, the diagnostic log gets blocked for several hundred milliseconds.
Conditions: This occurs when a PRE processor module switchover occurs with multiple line cards installed.
Workaround: There is no workaround.
•
Symptoms: The Cisco uBR10012 router is returning docsIf3CmtsCmRegStatusIPv6Addr as a string. This value should be in hexadecimal.
Conditions: This issue occurs if the modem does not have an IPv6 address. The default IPv6 addresses are displayed as a string.
Workaround: There is no workaround.
•
Symptoms: An incorrect MAC address or IP address is obtained when the show cable mode [mac | ip] access-group command is executed.
Conditions: This is seen in Cisco IOS Releases 12.3(23)BC and Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: The protect line card is showed as revertive in the show hccp detail command when no member subslot x/y revertive is configured.
Conditions: This issue is seen when line card redundancy is configured and after removing the working line card configuration and reconfiguring it.
Workaround: Reconfigure the no member subslot x/y revertive command.
•
Symptoms: Unintentional reset of the Cisco SIP-600 card can occur if the show controller command is executed and there is a Cisco Wideband SPA card that is currently in a state of initialization.
Conditions: This is seen if there is a SPA in a Cisco SIP-600 card that is in the process of being initialized.
Workaround: Do not execute the show controller command if there is a Cisco SIP-600 card with a SPA that is in the process of being initialized
•
Symptoms: The following traceback is seen when getting docsIfDownstreamChannelEntry:
SLOT 5/0: 00:09:18: %ALIGN-3-SPURIOUS: Spurious memory access made at 0x6061A388 reading 0x2B0SLOT 5/0: 00:09:18: %ALIGN-3-TRACE: -Traceback= 6061A388 60334794 6061D31C 606200D0 60519EBC 60519538 6050A97C 6050AB9CSLOT 5/0: 00:09:18: %ALIGN-3-TRACE: -Traceback= 6061A390 60334794 6061D31C 606200D0 60519EBC 60519538 6050A97C 6050AB9CSLOT 5/0: 00:09:18: %ALIGN-3-TRACE: -Traceback= 6061A3B4 60334794 6061D31C 606200D0 60519EBC 60519538 6050A97C 6050AB9CSLOT 5/0: 00:09:18: %ALIGN-3-TRACE: -Traceback= 6061A3C8 60334794 6061D31C 606200D0 60519EBC 60519538 6050A97C 6050AB9CConditions: This is seen when the broadband innovation upconverter is used.
Workaround: There is no workaround.
•
Symptoms: Frequency range overlapping occurs.
Conditions: This is seen when cable freq-range north-american/european is configured on the CMTS.
Workaround: Do not configure the cable freq-range north-american/european command in global mode.
•
Symptoms: After a PRE processor module failover, certain modems and their CPEs are not reachable by their IP addresses.
Conditions: This is seen in Cisco uBR10012 router with the ESR-PRE2.
Workaround: Fail over to original PRE processor module or execute the clear cable modem delete command on the affected modem.
•
Symptoms: Configuring more than four upstream channels in one downstream may cause a crash on the CMTS.
Conditions: This issue is seen while configuring more than four upstream channels in one downstream and when a UGS_AD service flow is used.
Workaround: Do not configure more than four upstream channels in one downstream.
•
Symptom: If Dynamic Message Integrity Check (DMIC) is configured in conjunction with IPv6 enabled CMs, the PRE will crash with the following message:
*** System received a Bus Error exception ***signal= 0xa, code= 0x8, context= 0x6493a1a4PC = 0x60a49ad0, Cause = 0x420, Status Reg = 0x34008002System Bootstrap, Version 12.0(20020314:211744) [REL-pulsar_sx.ios-rommon 112], DEVELOPMENT SOFTWARECopyright (c) 1994-2002 by cisco Systems, Inc.Reset Reason Register = RESET_REASON_RESET_REG (0x76)C10000 platform with 1044480 Kbytes of main memoryConditions: This issue is seen when DMIC is configured in conjunction with IPv6-enabled CMs.
•
Symptoms: The downstream load is not balanced when downstream load balancing group (DS LB) is configured with us-across-ds policy.
Conditions: This is seen when us-across-ds policy is configured on DS LB group.
Workaround: Do not configure us-across-ds policy on DS LB group.
•
Symptoms: An error "%GENERAL-2-CRITEVENT: MRI Unlink Error: Cable", is seen in the system logs for different CMs.
%GENERAL-2-CRITEVENT: MRI Unlink Error: Cable6/0/3 cpe_info->name: Cable8/1/3 current sid: 1464 cpe_info->sid: 1922current mri: 472 cpe_info->mri: 472modem->mac_addr: 001a.ad90.be36 cpe_info->mac_addr: 001a.ada7.c1a6Current IP address: 78.94.54.211 cpe_info->ipaddress: 10.80.64.206 cpe_info->next_slot: 1048576cpe_sidinstp: 0x0 cpe_sidinstp->prim_sid: 65535 cpe_sidinstp->cm_macaddr: 0000.0000.0000Conditions: This issue is seen on the Cisco uBR10012 router.
Workaround: There is no workaround.
•
Symptoms: The calculated channel utilization percents can become inaccurate. The short term utilization seen via the output of the show interface cable X/Y/Z mac-scheduler command, shows the "Avg upstream channel utilization" value much larger than the actual channel usage. The longer term utilization used by the load-balancing module may also be much larger than the actual channel usage.
Conditions: This issue is seen when "rate-adapt" is configured for a particular upstream channel under certain configured conditions. The MAC scheduler for that upstream allocates additional data grants to one or more cable modems in a given MAP message. When the data grants go unused by the cable modems, the utilization calculations made by the scheduler become skewed.
Workaround: There is no workaround.
•
Symptoms: The MAC Domain x/y/z interface loses modular-cable configuration.
Conditions: This is seen after a Cisco uBR100012 router bootup.
Workaround: There is no workaround.
•
Symptoms: A remote primary wideband cable modem sometimes forwards primary downstream service-flow traffic from a secondary wideband interface.
Conditions: This issue occurs in a remote primary wideband cable modem.
Workaround: There is no workaround.
•
Symptoms: The LED on the HCCP-protected downstream port lights up even when the protected downstream port is in standby mode.
Conditions: This is seen on a Cisco uBR10012 router with HCCP protect.
Workaround: There is no workaround.
•
Symptoms: After changing the load-interval on the RP, the interface load-interval on the line card does not change. This results in the utilization of load balancing updated using the default interval.
Conditions: This is seen while changing the load-interval on the legacy cable interface.
Workaround: There is no workaround.
•
Symptoms: An error message is displayed when the terminal monitor is enabled instead of displaying an error message when a terminal monitor is not present. The error message that needs to be displayed as follows:
000715: Feb 25 18:35:51.554: %Advanced Spectrum Management does not support three Modulation Profile configurations. Modulation Profile 221 has been removed from the Dynamic Modulation configuration.SLOT 5/0: Feb 25 18:35:51.606: %UBR10000-5-USFREQCHG: Interface Cable5/0/0 U0, changed to Freq 26.230 MHz Feb 25 18:35:51.606: %Advanced Spectrum Management does not support three Modulation Profile configurations. Modulation Profile 221 has been removed from the Dynamic Modulation configuration.Condition: This issue occurs when the upstream is configured for three-step dynamic modulation functionality and than upstream port is added to the spectrum group.
•
Symptoms: The cable metering commands are configured via SNMP as well using CISCO-CABLE-METERING-MIB objects. Because changes made via SNMP are not synchronized between the primary and secondary RPs, the cable metering-related configurations done via SNMP are not seen in the secondary PRE and configurations are lost after PRE switchover.
Conditions: This issue is seen with a high availability CMTS and images having CISCO-CABLE-METERING-MIB support.
Workaround: There is no workaround.
•
Symptoms: A wrong description on a facility alarm shows up for some upstream ports after resetting a modular-host line card. The wrong description means that "Physical Port Administrative State Down" shows up for some upstream ports even when those upstream ports are not administratively shut down.
Similarly, a wrong description on a facility alarm shows up for some upstream ports after reloading a Cisco uBR10012 router. The wrong description means that "Physical Port Link Down" shows up for some upstream ports even when those upstream ports are up.
Conditions: This issue only occurs on line cards with upstream max-ports 6 configured.
Workaround: Execute the shut command followed by a no shut command on the upstream ports.
•
Symptoms: A wideband CM cannot go into the w-online state after resetting both the working line card (guardian line card) and the protect line card and after the working line card reboots and becomes active.
Conditions: This issue is seen while resetting the working line card which is the guardian line card and when the working line card reboots first and becomes active.
Workaround: Remove and reconfigure the guardian configuration.
•
Symptoms: The PRE processor module crashes when the subinterface bundle is removed.
Conditions: This issue is seen when the CM has a lawful intercept configured and the CM is removed before the subinterface bundle is removed.
Workaround: Remove the bundle before removing the CM. For example, remove the subinterface bundle before shutting down the cable interface, or remove the subinterface bundle before executing the clear cable modem command.
•
Symptoms: Some DOCSIS 3.0 cable modems are not able to get to the w-online state.
Conditions: This is seen in DOCSIS 3.0 cable modems that require the presence of TLV7 and/or TLV8 in the primary channel MAC Domain Descriptor (MDD) in order to consider the MDD valid.
Workaround: There is no workaround.
•
Symptoms: The show cable modem remote-query command on the CMTS reports a modem downstream power level as a positive value while the SNMP query shows a negative value for the downstream power level between (-1 and 0).
Conditions: This issue is seen when "remote query" is enabled and the show cable modem remote-query command is issued. It only shows if the modem downstream power level is between 0 and -1.0 dBmV.
Workaround: There is no workaround.
•
Symptoms: The DOCSIS Timing, Communication and Control (DTCC) card reference source changes with execution of redundancy force-failover main-cpu command.
Conditions: This issue occurs during a PRE switchover on a Cisco uBR10012 router and the DTCC clock reference changes resulting in the following messages being shown:
%UBR10KTCC-4-CHG_CLK_REF: Clock reference source set to DTI for DTCC card 2/1 by process Plugin Start*Mar 3 08:26:03.172 CST: %UBR10KTCC-1-DTISLOT: DTCC slot 2/1: card role changed to Active*Mar 3 08:26:03.172 CST: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 2/1 port 1 DTI client status changed to Active normal*Mar 3 08:26:03.172 CST: %UBR10KTCC-1-DTIPORTSERVER: DTCC slot 2/1 port 1 DTI server status changed to freerun, external clock source: none*Mar 3 08:26:03.620 CST: %UBR10KTCC-4-CHG_CLK_REF: Clock reference source set to DTI for DTCC card 1/1 by process Plugin Start*Mar 3 08:26:04.172 CST: %UBR10KTCC-1-DTISLOT: DTCC slot 1/1: card role changed to Standby*Mar 3 08:26:05.172 CST: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active normal*Mar 3 08:26:05.172 CST: %UBR10KTCC-1-DTIPORTSERVER: DTCC slot 1/1 port 1 DTI server status changed to normal, external clock source: noneWorkaround: There is no workaround.
•
Symptoms: Cable metering commands can be configured via SNMP as well using CISCO-CABLE-METERING-MIB objects. Because changes made via SNMP is not synced between the primary and secondary RPs, cable metering-related configurations done via SNMP are not seen in the secondary PRE and configurations are lost after a PRE switchover.
Conditions: This issue is seen with a high availability CMTS and images having CISCO-CABLE-METERING-MIB support.
•
Symptoms: The wideband CM cannot be in the w-online state. It can only be online.
Conditions: This issue occurs when high availability synchronization is not completed.
Workaround: Reconfigure the fiber node.
•
Symptom: A traceback is seen during line card static synchronization.
Conditions: The traceback is seen occasionally when the length of data to be sent is near to the Maximum Transmission Unit (MTU).
Workaround: There is no workaround.
•
Symptoms: The RCC receive-module first-center-frequency encoding in the registration response message is not correct when using a rcc-template with multiple receive-modules. The CM rejects this INVALID-RCC error message in the registration ACK.
Conditions: This issue is seen in a DOCSIS 3.0 compliant CM using RCC-template which is configured with multiple receive-modules and distinct first-center-frequencies.
Workaround: Modify the RCC-template to use a single receive-module, but this may restrict the CM's capabilities.
•
Symptoms: Load-balancing of default routes originated by Open Shortest Path First (OSPF) from upper layer switches is not working.
Conditions: This issue is seen in a CMTS running Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: For the Cisco uBR-10-MC5X20 NB CM with the primary channel on remote a Cisco Wideband SPA downstream interface, its toaster BPI index is reset to 0 after a line card switchover; thus the downstream traffic is not encrypted.
Conditions: The condition is unknown.
Workaround: There is no workaround.
•
Symptoms: The device classification of certain types of CPEs impacts the cable filter groups applied to the CPE.
Conditions: This is seen in Cisco routers such as the Cisco 831 and Cisco 1841 used as CPEs behind CMs and are classified as "MTA" and not as "CPE". As a result, an unexpected cable filter group, or, in cases where defaults are not configured for "MTA", no filter group is applied to the device. A serious security gap is thus possible as all traffic from behind the device bypasses the established security policies.
•
Symptoms: Logging on to the console is disabled on the cable line cards.
Conditions: This issue is seen if any debugging is done. The logging must be explicitly enabled using the "logging console". Previously, the logging was enabled by default on cable line cards.
•
Symptoms: A preamble length of 49 symbols causes an issue with a Broadband BCM3300 chip with H cards. Imposing the minimum 50 symbols preamble length as to IM and SM bursts still reports the same issue.
Conditions:
Workaround: Do a shut command followed by a no shut command on the interface or execute cable upstream shutdown or a no cable upstream shutdown.
•
Symptoms: The Downstream Service Group (DS-SG) is missed after resetting the protect line card and working line card.
Conditions: This issue is seen while configuring Channel Grouping Domain (CGD) under c5/0/1 using the downstream modular-Cable 1/0/0 rf-channel 2 command and adding a modular interface to the fiber node.
Router#cable power off 5/1Router#cable power on 5/1sleep about 5 secondsRouter#cable power off 5/0Router#cable power on 5/0Workaround: Perform a line card switchover once more.
•
Symptoms: The show cable rf-status command is supposed to display the RF channels whose logical state shows "up", but sometime when the command is issued for individual channels separately, it displays RF channels that are not logically "up".
Conditions: This is seen while executing the show cable rf-status command.
Workaround: Use the show cable rf-status command on all the interfaces instead of trying to check each individual rf-channel separately.
•
Symptoms: The SNMP service primary flow counters include the pre-registration statistics while the corresponding commands do not include it. The ARP packets from CMTS are not mapped to the MAC destination address-based MAC classifiers on the Cisco uBR10012 platform.
Conditions: This happens for both upstream and downstream ports on the Cisco uBR7200 router and for the upstream port on the Cisco uBR10012 router.
Workaround: There is no workaround.
•
Symptom: When the CMTS receives IGMPv3 SSM Leave, the multicast session may time out prematurely.
Conditions: This issue is seen when IGMPv3 Leave is received by the CMTS.
•
Symptoms: Bonding-capable CMs may take a long time to get to the w-online state, or possibly fail to get to the w-online state.
Conditions: This issue is seen when the CMTS is configured to forward for a bonding-capable CM to a bonding-capable primary downstream channel, the CMTS randomly selects the target from the set of DS channels defined in the MAC Domain Downstream Service Group (MD-DS-SG). When complex fiber nodes are configured, it is possible that one or more of those channels are primary in a different DOCSIS MAC domain. In the event the upstream channels for that MAC domain are not available at the current fiber node, then the CM fails to initialize at the target frequency. Due to the random nature of the target DS selection, multiple failures are possible.
Workaround: There is no workaround.
•
Symptoms: Multicast Downstream Service Identifier (DSID)s are not synced over.
Conditions: This issue occurs when there is no hardware or the hardware is not up in the line card slot, before a line card that hosts modular channels with multicast traffic.
Workaround: There is no workaround.
•
Symptoms: In a PRE processor module switchover, the standby is reset repeatedly by the active because of a mismatch in the command list.
Conditions: This is seen while configuring IPDR sessions, and the metering session, and then removing the IPDR session.
Workaround: Configure the IPDR corresponding session as in the below example and follow the procedure to unconfigure the cable metering session with IPDR if it is already configured.
–
Router(config)# cable metering ipdr 2 session 100–
Router# show running-config | include cable metering ipdrRouter# cable metering ipdr 2 session 0–
Router(config)# no cable metering ipdr 2 session 100•
Symptoms: The CMTS exhausts more than 50 percent of the CPU utilization with "getmany docsIf3MdUsToDsChMappingTable".
bash-3.00$ for ((i=0; i<100; i++)) do ` getmany -v2c 20.4.0.48 public docsIf3MdUsToDsChMappingTable>/dev/null`; done;nm-ubr10k-3#sh process cpu | i SNMP ENGINECPU utilization for five seconds: 0%/0%; one minute: 0%; five minutes: 2%PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process346 84140 57115 1473 67.11% 45.98% 14.07% 0 SNMP ENGINE ========> HighCondition: This is seen when an SNMP get operation is performed.
•
Symptoms: A traceback is seen when the show pxf cpu queue hqf pblt detail command is executed.
Conditions: This is seen when pblt_id is invalid and the according pblt is not used.
Workaround: Use the correct pblt_id.
•
Symptoms: The modem gets stuck in the init (io) state when both working and protect line cards are reset at almost the same time.
Conditions: This issue is seen when both the working and protect line cards are reset.
Workaround: Reset the working card using the hw-module subslot x/y command or do a switchover.
•
Symptom: The CMTS can crash.
Conditions: The issue is seen when ETDB show commands are paused and multicast sessions are removed because of a CPE leave or a CM goes offline.
Workaround: The term length can be set to 0 for no pausing during command display.
•
Symptoms: During ISSU commit version for the Cisco uBR10012 router, these errors are seen:
00:27:58: %COMP_MATRIX-3-PRST_VBL_DELETE: ISSU Compatibility Matrix was unable todelete a persistent variable from the peer system (rc = -2).[OK]00:30:19: %COMP_MATRIX-3-PRST_VBL_WRITE: ISSU Compatibility Matrix was unable towrite to persistent variables on the peer system (rc = 0).Conditions: This issue is seen during ISSU commit version for the Cisco uBR10012 router, and while upgrading from Cisco IOS Release 12.2(33)SCB2 to Cisco IOS Release 12.2(33)SCB3 or later.
Workaround: There is no workaround.
•
Symptoms: A configuration mismatch is seen between card slot and card type while configuring a card type. The configuration failed and an error message is displayed.
Conditions: This is seen on the Cisco uBR10012 router.
Workaround: Configure the correct card type for the card slot.
•
Symptoms: CM fails to get online.
Conditions: This issue is seen when a CM begins initial-ranging on an upstream port, and reports a DOCID for a different MAC domain.
The CMTS issues a downstream frequency override (DFO) to the downstream frequency of the hosting interface (for example, Cisco uBR10-MC5X20 line card) of the upstream port being used. In this event, the target downstream port is in an "rf-shutdown' state", and the DFO fails.
Workaround: There is no workaround.
•
Symptoms: The following syslog is generated at a very high rate:
%UBR10000-6-MAP_UPSTREAM_BW_ALLOCATION:This can cause high CPU utilization due to an interrupt on a Cisco uBR10012 router.
Conditions: The condition is unknown.
Workaround: Disable console logging or raise the minimum level sent to the console.
•
Symptoms: The upstream connector configuration is different when the show run command is executed on the active PRE and standby PRE processor modules.
Conditions: This issue is seen in a line card switchover.
Workaround: Perform a line card revertback and a line card switchover again.
•
Symptoms: The SA Descriptor Length field size is changed by 2 bytes.
Conditions: This update is required by EuroDOCSIS standard (ECR).
•
Symptoms: When sending Dynamic Bonding Change (DBC) to a CM, while adding or deleting a host on the MAC domain, the CMTS always sends with a GMAC authentication which is not required for Main Distribution Frame (MDF) 2 CM. The "Re-seq" flag is set to 0, which causes multicast forwarding to stop at the CM.
•
Symptoms: The cable metering options "flow-aggregate" and "cpe-list-suppress" get lost from the cable metering configuration if cable metering data-per-session x timer y is present.
Condition: This is seen in the cable metering command.
Workaround: Reconfigure the metering options "flow-aggregate" and "cpe-list-suppress" in the command.
•
Symptoms: After the Cisco uBR10012 router is upgraded from Cisco IOS Release 12.3(13a)BC1 to Cisco IOS Release 12.3.(23)BC2, the advanced spectrum management feature does not work.
Conditions: The SNR value goes below the first threshold and enters the second modulation as required. The issue occurs if a change is made to the first modulation, using the second threshold as reference. For example:
Router# cable upstream 0 threshold snr-profiles 22 5Mar 5 17:13:48 M 16-QAM QPSK SNR 21<22 CNR 21<22 CFEC 11>=10 <<CORRECTMar 5 17:14:10 M QPSK 16-QAM SNR 24>=8 CFEC 0<=10 UnCFEC 0<=1 << WRONG, it \should be SNR >=25Workaround: Disable the second threshold value cable upstream 0 threshold snr-profiles 22 0 command.
•
Symptoms: The calculated upstream port utilization percentage on ports 1 to 3 is below the actual usage.
Conditions: This is seen when more than one upstream port is active for a given MAC domain. The upstream utilization for the higher port numbers can be too low.
Workaround: There is no workaround.
•
Symptoms: Enhanced Interior Gateway Routing Protocol (EIGRP) reports a bad checksum.
Conditions: This is seen when EIGRP adjacency is configured over a Generic Routing Encapsulation (GRE) tunnel.
Workaround: Disable PXF on the Cisco uBR10012 router.
•
Symptoms: When a show diagnostic command is executed, the card uptimes are reported as a factor of weeks, while others commands display the card uptimes as a factor of days such as "Card uptime 12 days, 7 hours, 26 minutes, 12 seconds" instead of "Card uptime 1 weeks, 5 days, 7 hours, 25 minutes, 45 seconds".
Conditions: This is seen on a Cisco uBR10012 router running Cisco IOS Release12.2(33)SCB2.
•
Symptoms: Manually setting the MTU to "mtu 1500" on interface modular-cable is not reflected in the show interface command.
Conditions: This issue is seen on a Cisco uBR10012 router running Cisco IOS Release 12.2(33)SBC.
Workaround: First manually set the MTU to "mtu 1464" and then manually set it to "mtu 1500".
•
Symptoms: Traceback occurs when the show pxf cpu statistics ip-session command is executed.
Conditions: This traceback is always shown whenever these commands are executed.
Workaround: There is no workaround.
•
Symptoms: A Cisco uBR10012 router running on Cisco IOS Release 12.2(33)SCB1 crashes with following error:
"PXF DMA Toaster Stall Error".
Conditions: The crash happens during normal operation and is caused by a corrupted packet.
Workaround: There is no workaround.
•
Symptoms: SNMP query docsIf3MdChCfgTable is much slower than "ifStackTable".
Conditions: This is an issue in the SNMP walk of docsIf3MdChCfgTable.
Workaround: There is no workaround.
•
Symptoms: Some wideband interface association with bundle-related errors may occur if line card switchover and revertback is done once on active the PRE-A and the same action is performed on PRE-B after a PRE switchover to PRE-B.
Conditions: This issue occurs on line card switchover.
Workaround: There is no workaround.
•
Symptoms: The DHCPv6 relay process crashes on the active RP.
Conditions: The condition is unknown.
Workaround: There is no workaround.
•
Symptoms: The critical facility alarms are incorrectly reported for wideband SPA SFPs.
Conditions: This issue is seen on a Cisco uBR10012 router running Cisco IOS Release 12.2(33)SCB.
Workaround: There is no workaround.
•
Symptoms: The standby PRE module reloads if cable metering is unconfigured while metering is in progress.
Conditions: This is seen if cable metering is unconfigured.
Workaround: Ensure metering is completed and not in progress using the show cable metering status verbose command before unconfiguring the cable metering command.
•
Symptoms: When SF-based assignment is used to assign SF forwarding to a non-primary MC interface for the wideband CM, the downstream traffic fails.
Conditions: This is seen when attribute-based forwarding on a wideband CM attempts to assign forwarding to MC interfaces that are not the CM primary downstream interface.
Workaround: If forwarding is required to be non-bonded forward on the CM primary downstream interface, use the default bonded forwarding.
•
Symptoms: All the 65535 Blaze indices are used up.
Conditions: This happens when the guardian is remote and there are multiple service flows per CM.
Workaround: Use a local guardian.
•
Symptoms: The ISR processes and higher priority tasks on the line card are busy so that IPC keepalive messages cannot get disposed in time, which could lead to the IPC connection going down and a line card crash.
Conditions: This issue is seen in a large environment, for example, over 3000 modems in one line card. During startup, the line card may crash due to an IPC timeout.
Workaround: There is no workaround.
•
Symptoms: Resetting an IPv6-capable CM on a CMTS with the cable interface configured for "cable ip-init ipv6" when it previously came online on IPv4 results in both IPv4 and IPv6 addresses and setting the DIP to "Y".
Conditions: This is seen in Cisco IOS Release 12.2(33)SCB.
Workaround: This issue does not cause any operational impact at all. As a best practice, when changing IPv4 to IPv6 only, use the clear cable modem xxxxx delete command rather than a reset. A delete will clear this issue.
•
Symptoms: Excessive numbers of DOCSIS DSx message failures are seen, including DSA_ABORT_NO_ACK, DSC_ABORT_NO_ACK.
Conditions: This is seen in an environment that includes a significant amount of call signalling. The problem is exacerbated by a significant amount of IOS process activity by SNMP.
Workaround: Decrease the SNMP activity by applying SNMP blocks on the RP.
•
Symptom: An enhancement in the ISSU packets size correctly for large messages and avoid a memory leak.
Conditions: This is an enhancement in the ISSU packet size.
Workaround: There is no workaround.
•
Symptoms: Downstream service group information disappears.
Conditions: This issue is seen while powering off the working line card, and when the shutdown command followed by a no shutdown command is executed on the interface in a protect line card.
Workaround: Do not execute a shutdown command followed by a no shutdown command on a protect interface when the working line card is powered off.
•
Symptoms: The IF-MIB ifSpeed for the local downstreams show 42.8 Mbps throughput on CMTS running Cisco IOS Release 12.2(33)SCB2. This should be brought down to 37.5 Mbps to better account for the overhead presented by FEC+Trellis+MPEG2 overhead.
Conditions: This is seen in Cisco IOS Release 12.2(33)SCB2.
•
Symptoms: Online insertion and removal (OIR) compatibility fails when it tries to reconfigure the MAC domain with SPA channels 16-23 on the new card.
Conditions: This is seen on OIR replacement of the Cisco MC5X20 cards. The cleanup of SPA channels 16-23 does not happen properly.
Workaround: There is no workaround.
•
Symptom: The following message is observed after a PRE processor module switchover:
00:02:51: %UBR10KTCC-1-DTIERR: DTI interface Error code -7 function dti_set_active_cardConditions: This issue is seen when there are two TCC cards.
Workaround: There is no workaround.
•
Symptoms: Removing all 802.1Q VLANs on an EtherChannel crashes PXF on the PRE4 processor module, if the keepalive is enabled on any Gigabit Ethernet interface, causing a "PXF DMA FTC Unknown Command Error". This only happens for SPAs on Cisco 10000 Series SPA Interface Processor-600 (Cisco SIP-600).
After system reload, the show pxf cpu ech port-channel xx.yy commands fails to display the hash table of VLANs on both the PRE2 and the PRE4.
Conditions: This is seen when the EtherChannel "cicb_data.ech_bundle_index" on External Column Memory (XCM) column 5 is set to zero when the native VLAN, vlan1, is destroyed. As the result, a keepalive packet is incorrectly sent via the VCCI of the port-channel, instead of the Gigabit Ethernet interface.
The second symptom is due to the failure to update 802.1Q the vlan's "cicb_data.ech_bundle_index" on XCM column 5 when an interface is added to an EtherChannel.
Workaround: Restore the traffic in four steps:
1.
2.
3.
4.
In theory, you may boot the CMTS with no EtherChannel configured, and copy an edited production configuration file to running-config after all the interfaces are up.
•
Symptoms: When setting a CA certificate to "untrusted", any CM that uses an issuer of the same name is rejected, including the legitimate modems.
Conditions: The issue is found because of a newly created software "Haxorware" which generates these CA certificates that conflict with the existing CA certificates.
Workaround: The recommended method is always to not allow self-signed certificates on the CMTS and explicitly set specific self-signed certificates to trusted. This is the "opt-in" model, rather than the "opt-out" model.
•
Symptoms: IGMPv3 join is ignored by CMTS.
Conditions: This issue is seen when IGMPv3 is configured on the bundle interface. SSM-mapping is configured on the CMTS. The ignored IGMPv3 join is *,G join with a group address in the SSM range.
Workaround: Use S,G IGMPv3 join or use *,G IGMPv2 join.
•
Symptoms: The Channel Grouping Domain (CGD) information is not displayed in the show run command or the show cable mac-domain cgd-associations command when no connector is configured.
Conditions: This issue is seen when no connector is configured for the upstream port.
Workaround: Configure the connector for the upstream port.
•
Symptom: Modem registration is not disabled when a guardian card is missing.
Conditions: This is seen in the configuration of the online CM remote downstream interface.
Workaround: Configure the guardian line card using the modular-host command or configure the CM online on a local Cisco uBR10-MC5X20 downstream line card.
•
Symptoms: Modems using SPA modular-cable downstreams on the Cisco uBR10012 router cannot go past the init (io) state.
Conditions: This is seen when multicast encryption is enabled. This issue occurs when a key index request for an encrypted multicast session from the MAC domain host line card to the wideband host line card fails. Such a failure can occur due to an inter-line card IPC drop.
Workaround: The following steps can be used to recover from this error state:
1.
2.
3.
•
Symptoms: When the wideband resiliency is not configured, output for the show cable modem summary wb-rfs command is null or not correct.
Conditions: This is seen if the CMTS never configures wideband resiliency and if wideband resiliency configuration is removed from the CMTS, then the output is incorrect.
Workaround: There is no workaround.
•
Symptoms: During PRE module switchover, each of the line cards unnecessarily updates the toaster information.
Conditions: This issue is seen during a PRE switchover triggering the update.
Workaround: There is no workaround.
•
Symptoms: The Gigabit Ethernet output for a 24 downstream wideband and narrowband SPA can getisolated from the port after repeated online insertion and removal (OIR) of the SPA within a short duration of time.
Conditions: This issue is seen with repeated OIR of the SPA within a short duration of time and with repeated line protocol off and on within a short duration of time.
Workaround: Reload the SPA using the hw-module bay reload command.
•
Symptoms: When using a wideband SIP, the total count for the SPA in slot 1/1/0 is a sum of both SPAs, rather than the sum of itself. This can be found using the show hw-module bay all counters rf-channel command.
Conditions: This issue is seen in a wideband SIP when more than one SPA is inserted in one SIP.
Workaround: There is no workaround.
•
Symptoms: Per MAC domain statistics are not available.
Conditions: This issue is seen when debugging for forwarding statistics per MAC domain using multiple commands.
•
Symptoms: The failure of a DOCSIS SPA with modular cable (MC) interfaces occurs when there is multicast traffic.
Conditions: This happens when multicast traffic is forwarding to a MC interface with multicast QoS configured and a hw module shutdown on SPA causes the RP to crash.
Workaround: There is no workaround.
•
Symptoms: The standby PRE processor module keeps on reloading with a bulk-sync failure when the Cisco Full Height Gigabit Ethernet (FHGE) card is configured in the slot.
Conditions: This issue in the preconfiguration of the FHGE in the slot.
Workaround: Do not use FHGE in slot 2.
•
Symptoms: A channel enters the disabled state in a load balancing (LB) state machine.
Conditions: This issue occurs if there is a channel whose condition is very bad, and is configured to a LB group with other channels. The state is set to disabled if it has too many LB failures on it. If this bad condition persists for a long time, and then it reverts, the LB state of this channel may get disabled.
Workaround: Manually execute the clear cable load-balance state command .
•
Symptoms: After a cable line card switchover, the multicast traffic is forwarded by the CMTS but the traffic does not reach the customer premises equipment (CPE) behind the cable modem.
Conditions: This issue is seen when a multicast QoS group is configured with encryption and there is a session using that QoS. This issue is seen after line card switchover for that multicast session.
Workaround: Send an IGMP Leave from the CPE and then let the CPE join the session again.
•
Symptoms: The Cisco SIP-600 removal of a wideband SPA from bay 2 affects traffic on the wideband SPA in bay 0. The affected service flow queues never dequeue packets resulting in a full queue and taildrops.
Conditions: This is seen in the Cisco SIP-600 in slot 1 or 3 and wideband SPAs in bays 0 and 2 of the SIP.
Workaround: Perform online insertion and removal (OIR) or power cycle the affected SPA or SIP.
•
Symptoms: The Cisco uBR10012 router may suffer a PXF crash with the following message in the PXF crash information file:
PXF DMA Error - End of Descriptor Before Cmd Byte Length ExhaustedConditions: The crash is observed on the Cisco uBR10012 router with PRE2 processor module running Cisco IOS Release 12.2(33)SCB2.
Workaround: There is no workaround.
•
Symptoms: The SPA BPI key index for a multicast Security Association Identifier (SAID) shows as zero after a PRE module or line card switchover. The index is displayed as Blaze_index in the show interface cable <x/y/z> key sid command.
Conditions: This is seen after PRE module or line card switchover.
Workaround: Remove the cable match command in the bundle interface and reset all the line cards that have multicast SAIDs with 0 SPA BPI index. If line card high availability is configured, keep the protect line card down.
•
Symptom: The Cisco uBR10012 router with a PRE2 processor module configured for PacketCable Multimedia (PCMM) may crash with the following message in the log:
Invalid SID (1004) position for interface Cable6/0/4: CM 0015.a4f8.2076:Is used by CM <mac> SFID 17066 SID 120. SID container info: start 7933 end 7275-Traceback= 608F7634 608F7D50 60221758 60221CC4 60222A48 6035198C 60302754 60303B98 603040C0 6034EACC 609F1870 609F185CConditions: The crash is observed on the Cisco uBR10012 router with a PRE2 processor module running Cisco IOS Release 12.2(33)SCB2 and with PCMM configured.
Workaround: There is no workaround.
•
Symptoms: A memory leak is observed when unconfiguring the cable downstream dsg tg channel command from the interface.
Conditions: This was seen when cable downstream dsg tg channel is unconfigured.
•
Symptoms: Multiple memory leaks are observed while issuing the show memory debug leak command on the line card.
Condition: The condition is unknown.
Workaround: There is no workaround.
•
Symptoms: PXF crashes after "microcode reload pxf".
Conditions: This issue is seen when the following conditions are done simultaneously:
–
–
–
Workaround: To prevent a crash from happening:
1.
2.
•
Symptoms: A crash is observed when running the show tech | redirect disk0:xxxx command.
Conditions: The following conditions may cause a crash:
–
–
Workaround: The crash does not occur with a Telnet session. If you use SSH to connect into the CMTS, do not redirect the output to DISK0 or DISK1.
•
Symptoms: The CMTS fails to send Subscriber Account Management Interface Specification (SAMIS) data to the IPDR collector, though the TCP connection is still there. The show ipdr session command reports there is unacknowledged data, and show process event of IPDR process reports the process is in "sleep" state.
Conditions: IPDR and cable metering is enabled.
Workaround: Disable IPDR and re-enable.
•
Symptoms: A traceback is displayed with the "IPCTXERR" error message.
Conditions: This is seen when IPC fails due to various reasons and the "IPCTXERR" error message is displayed with traceback.
Workaround: There is no workaround.
•
Symptoms: Excessive numbers of DOCSIS DSx message failures are displayed, including DSA_ABORT_NO_ACK, DSC_ABORT_NO_ACK.
Conditions: This issue is seen in the call signaling environment and by a large amount of IOS process activity by SNMP.
Workaround: Decrease the SNMP activity by applying SNMP blocks on the RP.
Open Caveats—Cisco IOS Release 12.2(33)SCB2
•
Symptoms: Swapping unlike the Cisco uBR10-MC5X20S/U line cards ) causes modems to go offline, and the configuration is lost.
Conditions: The default behavior of the cr10k card [slot/subslot]; oir-compatibility command must be converted from default disabled to default enabled for all cable line cards.
Workaround: Prior to exchanging line cards, configure OIR compatibility for all slots. If line card exchange happens without configuring OIR compatability, and the problem is discovered before a write memory is issued, perform the following workaround:
1.
2.
3.
4.
This is the only procedure that can ensure that the frequency stacking and virtual interface configurations are preserved. Attempts to paste pieces of the previously stored running configuration will fail if frequency stacking or virtual interfaces are configured, because the connectors must be unassigned first.
If a write memory has occurred, then the shutdown state and blank configuration for all interfaces will be written to both primary and secondary NVRAM.This technique will not work without resorting to an externally stored backup configuration for the system.
•
Symptoms: The throughput rate shows (slightly) more than the expected range.
Conditions: This happens when the the DMIC feature is enabled and the modem is marked with given maximum rate values. This can also happen when the QoS profile has lower maximum rates.
Workaround: There are no workarounds.
•
Symptoms: The number of downstream PHS rules for a cable MAC domain stays the same even though more downstreams can be added to the MAC domain (through modular downstreams).
Workaround: There are no workarounds.
•
Symptoms: When a UCD change occurs, some data structures related to the US scheduler of secondary upstream service flows do not get properly updated (for example, minislot size). Because of this, any traffic through these service flow might get affected (for example, packet drop).
Workaround: There are no workarounds.
•
Symptoms: Multicast stream drops after changing ip igmp query-interval
Conditions:
–
–
–
Workaround: There are no workarounds.
•
Symptoms: 12.3(23)BC2 mixed docsis mode (atdma-tdma) can only achieve 40 UGS for an upstream. But, 12.3(17b)BC9 can achieve 48 UGS flows.
Conditions: Currently, it looks like this may have something to do with mixed mode and modems.
Workaround: There is no workaround.
•
Symptoms: Inactive sessions are observed in the multicast explicit tracking database.
Conditions: IGMP leave received for a group followed by a cable line card switchover
Workaround: There are no workarounds.
•
Symptoms: Cable metering collection enters "write-error" state and does not recover.
Conditions: When cable metering is configured with default TCP parameters.
Workaround: Enabling "ip tcp path-mtu-discovery" may help prevent occurrences of the issue. Also, the command test cable metering abort will clear the "hung" state and allows the next iteration of cable metering to occur.
•
Symptoms: The cable line card crashes after the execution of the commands clear arp and clear cable modem all delete.
Workaround: There are no workarounds.
•
Symptoms: Certain PCMM calls using wideband CMs may have one way audio.
Conditions: PCMM calls between w-online(pt) and online(pt) on wideband CMs.
Workaround: There are no workarounds.
•
Symptoms: When MVPN is configured on CMTS, and a CPE joins a multicast group that has a source already transmitting to, then the (S,G) entry is not created in mvrf table of the CMTS.
Workaround: There are no workarounds.
•
Symptoms: When modem comes up during the off-peak time, when it is not supposed to be monitored as per enforce-rule configured, the smp is not synced to standby line card.
Conditions: If a line card switchover is done during the off-peak time, even if the modem enters peak-time, it is not monitored as per the enforced rule configured. The smp is synced to standby line card only once the modem is monitored during peak-time.
•
Symptoms: In numerous 10Ks running 12.3(21a)BC7 and BC8, the customer is seeing instances where the CPE device disappears from the cable host table and sometimes from the ARP table. In all cases, the CPE loses IP connectivity. This shows up across different CM vendors, different CLC types, different plants, and different 10Ks.
Conditions:
–
–
–
–
Workaround: Try clear cable host until the device appears back in the host table. This may not work always.
•
Symptoms: Secondary PRE crashes when DBS function is applied to wideband interface.
Conditions: Problem is seen on a PRE-2 running 12.3(23)BC2. To apply DBS, the customer has to shut down the wideband interface first and then they met a IOS crash when "chunk_free(docsis_srv_flow_chunk,flowp)" at the last step of cmts_destroy_ds_service_flow on secondary PRE. No PXF happens.
Workaround: There is no workaround.
•
Symptoms: In a very rare scenario, after an N+1 switchover, upstream traffic that is BPI encrypted is not received properly by the CMTS. Instead, input errors are logged on the interface and "debug cable error" gives error messages similar to:
Cable5/1/4: Bad rx packet. JIB status code 0xA
Conditions: The issue occurs on upstream channels that use a "shared" connector where the other upstream channel using the same shared connector is on another downstream and is shutdown.
Workaround: Activate the downstream and other upstream channel using the same shared connector or unshare the upstream connector temporarily.
•
Symptoms: The show controller modular-cable 1/0/0: command shows that rf-channel 0 become unknown after hw-module reset.
SPA IP address = 192.168.5.1 SPA MAC Addr = 0012.001A.67D7 QAM MOD ANNEX TKB Interval Rate adjust State 0 Unknown Unknown 0 1 Enabled ====> unknown 1 QAM 256 Annex B 2423 132 Enabled 2 QAM 256 Annex B 2423 132 Enabled 3 QAM 256 Annex B 2423 132 Enabled
Conditions: Happened after hw-module reset and show diag 1/0/0. This will cause all wideband modems also go to NB.
Workaround: Reload the router to correct the issue.
•
Symptoms: On a UBR10k the input rate counter of all GRE tunnels is incorrect. It is doubled compared to the real value. Only tunnels interfaces are affected.
Conditions: This occurs on a UBR10k/ESR-PRE2 running IOS versions 12.3(23)BC1 and 12.3(23)BC2 and 12.3(23)BC5.
Workaround: There is no workaround.
•
Symptoms: From logs collected from the customer's network, at least three routers show the symptom of the problem. For the two SPAs installed in the Cisco Wideband SIP jacket card, SPA1/0/1 lost all its configurations, including its IP and MAC addresses, and all its RF channel configurations. SPA1/0/0 lost some of its configurations. IP and MAC addresses remain but some of the RF channel configurations are missing. Recovery measures were attempted, including PRE S/O, hw-module reset, unprovision/reprovision the SPAs, but none succeeded. Reloading the router seems to be the only way to recover.
Workaround: Reloading the router is the only way to recover.
•
Symptoms: PRE HA test shows that sometimes the voice call dropout time is larger than one second.
Conditions: Test the offnet call.
Workaround: There are no workarounds.
•
Symptoms: Modems on modular interface suffer long-time recovery time or even fall offline after switchover or revertback.
Conditions: Only affect 520S/520U card as active card. Only affect modems on modular interface. More visible for annex A modulation type.
Workaround: There is no workaround.
•
Symptoms: For deleted service flows the last character in the Service Class Name field is dropped in SAMIS records as well as in the SNMP MIB docsQosServiceFlowLogServiceClassName.
Workaround: There are no workarounds.
•
Symptoms: Memory leak at CMTS_HCCP_GET_TLV_PARSE_TLV has been observed at the RP.
Workaround: There is no workaround.
•
Symptoms: In case of multicast and the forwarding is through WB interface, multicast SFs are created and the counters are maintained under the WB interfaces.
Workaround: There are no workarounds.
•
Symptoms: Memory leak at buffer_peak_notify is observed.
Workaround: There are no workarounds.
•
Symptoms: When TWC, the serviceflow in admitted state still has traffic going through.
Conditions: In three-way calling, and one side is on hold.
Workaround: There is no workaround.
•
Symptoms: SP messages are logged in RP console but not sent as traps by RP's SNMP agent.
Workaround: There are no workarounds.
•
Symptoms: Both the working cards 5/0 and protect card 5/1 are active.
Conditions: Do a LC SO from 5/0 to 5/1. After the synced, shut down 5/0/2 interface. Do a LC revertback from 5/1 to 5/0. Do no shut 5/0/2 (Do not wait in between step3 and step 4). Now you will find, after suspend timer expires, both 5/0 and 5/1 are active.
Workaround: Do not shut down an active protect interface. If both the cards (working and protect) are active, reset protect card.
•
Symptoms: Backup DTCC card may experience an unexpected reload irregularly with following messages:
Oct 14 23:01:25.487: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:00:00:07 LKA:00:00:07 LRMV:never, NKA:00:00:00, cpu 1%/1% 2%
Oct 14 23:01:25.491: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:never LKA:never LRMV:never, NKA:00:00:00, cpu 1%/1% 2%
Oct 14 23:01:25.491: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is down. Notifying 2cable-dtcc driver.
Oct 14 23:01:25.491: %UBR10K_ALARM-6-INFO: ASSERT CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1
Oct 14 23:01:26.555: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:26.555: %IPCOIR-5-CARD_LOADING: Loading card in slot 1/1
Oct 14 23:01:27.767: %C10K-5-LC_NOTICE: Slot[1/1] Line-card Image Downloaded...Booting... Oct 14 23:01:45.580: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:45.656: %UBR10KTCC-4-CHG_CLK_REF: Clock reference source set to DTI for DTCC card 1/1
Oct 14 23:01:45.936: %UBR10K_ALARM-6-INFO: CLEAR CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1
Oct 14 23:01:45.936: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is up. Notifying 2cable-dtcc driver.
Oct 14 23:01:47.040: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active fast
Oct 14 23:02:20.042: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active normalConditions: This problem is seen on only Backup DTCC irregularly of UBR10K running 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: Cable metering fails and enters a hung state.
Conditions: This problem happens when the ip tcp timestamp option is configured globally.
Workaround: Do not use the ip tcp timestamp option.
•
Symptoms: During LCHA, the CM instance bpi_enable flag (cminstp->priv_en) on RP is not synced over correctly. This is observed for MC interface CM with bighorn image.
Conditions: Generate multicast stream to CMs, bpi CM during LC switchover.
Workaround: There is no workaround.
•
Symptoms: MDD change message is generated on all the RF channels globally whenever any RF channel is configured on the CMTS. MDD change is supposed to be generated for channels within a MacDomain.
Workaround: There are no workarounds.
•
Symptoms: SNMP GetNext requests for docsQosServiceFlowPrimary (also known as 1.3.6.1.2.1.10.127.7.1.3.1.8 or docsQosServiceFlowEntry.8) are rejected. Still, if a certain docsQosServiceFlowPrimary entry is polled with SNMP Get directly (after some additional calculations are performed to determine the index value), the value is returned as expected.
Conditions: UBR7114E running 12.3(21a)BC3.
Workaround: Poll the individual values following the steps of the procedure suggested in SR 610144513.
•
Symptoms: Show interface cable x/y/z SID will show null MAC address and null IP address SID entries.
Conditions: IOS 12.3(17b)BC9, UBR10K CMTS
Workaround: There is no workaround.
•
Symptoms: When a UBR10k has a lots of CMes, the command show cable modem docsis version summary total may not be able to display each column properly because there is not enough space between columns.
Conditions: When there are lots of cable modems.
Workaround: There are no workarounds.
•
Symptoms: When upgrading IOS and hardware, the existing configuration could conflict with the channel IDs on other interfaces.
Conditions: Customer would like to see the DS channel IDs hidden, but viewable with a show command, in order to avoid potential conflicts.
Workaround: Reconfigure downstream channel-id for any interface that has a conflicting ID.
•
Symptoms: When either Cisco5x20 card Mac domains have WB associations, a traceback may be observed along with a message similar to the following when NO CARD x/x is used:
*Jan 25 05:41:57.235 PST: %SYS-3-BADLIST_DESTROY: Removed a non-empty list(2F3F7BF4, name: (unknown)), having 36 elements -Process= "C10K Card Actions", ipl= 0, pid= 75A small amount of memory leak may also happen. No other side effects are expected.
Conditions: This traceback is seen only if the card had a NB-WB associations previously, and was powered down or removed.
Workaround: There are no workarounds.
•
Symptoms: A Cisco router may reload reporting a bus error.
Workaround: There is no workaround.
Resolved Caveats — Cisco IOS Release 12.2(33)SCB2
•
Symptom: Multilayer Switch Feature Card (MSFC) crashes with RedZone memory corruption.
Conditions: This problem is seen when processing an auto-RP packet and Network Address Translation (NAT) is enabled.
Workaround: There is no workaround.
•
Symptom: A router may write a crashinfo that lacks the normal command logs, crash traceback, crash context, or memory dumps.
Conditions: This might be seen in a memory corruption crash depending on precisely how the memory was corrupted.
Workaround: There is no workaround.
•
Symptom: An unexpected reload occurred in Cisco SIP-600 carrier cards in Cisco UBR10012 router.
Conditions: This issue occurs when it is used with DS-24 SPA for DOCSIS 3.0
Workaround: There is no workaround.
•
Symptom: A change was introduced in Cisco IOS Release 12.3(21)BC that forces specific cable modems to be compliant to the DOCSIS specification. The requirement forces DOCSIS 1.1 modems to use BPI+ for authentication purposes instead of BPI. A number of customers have 10000s of cable modems that are not compliant as they do not contain CM certificates. Upgrading to this release is causing them a lot of pain and they are not willing to upgrade their modems at this time as it is cost prohibitive.
Conditions: The default behavior of the CMTS was changed in 12.3(21)BC with no way to disable the new default behavior.
Workaround: A workaround is to provision these modems to sign on using DOCSIS 1.0 configuration files which will make them DOCSIS 1.0 modems and compliant when using only BPI.
•
Symptom: The active PRE module crashes during bulk sync processing when the standby PRE module boots up in a Cisco uBR10000 router.
Conditions: This crash occurs when there is an aggregate of more than 400 active multicast replications on the SPA downstream interfaces (modular-cable or wideband-cable) during the PRE module bulk sync process. However, if the standby PRE module boots up before the SPA interfaces are up, the crash is not be seen since the session information is synced to standby PRE module using incremental syncs instead of bulk sync.
The multicast replications may be created either through static multicast configuration (possibly for DOCSIS Set-Top Gateway (DSG)) or for dynamic multicast sessions.
Workaround: Unconfigure the ip igmp static-group command and the ip multicast-routing command.
•
Symptom: The SPA interface went down with following error message in the log: "SPAWBCMTS,2,CARD_UP_DOWN,Card 24rfchannel-spa-1 in bay 1/0/0 is down., *C10K,3,LC_ERR,"Slot[1/0] 2jacket-1 WIDEBAND DOCSIS SPA[1/0] voltage outside of nominal range. Nominal (1800mV), value(2017mV) The SPA has been,*"
Conditions: This issue is seen in a Cisco uBR10000 router using a SPA-24XDS-SFP with PRE2 module running Cisco IOS Release12.3(23)BC6.
Workaround: Use hw-module subslot 1/0 reset command to recover the interfaces.
•
Symptom: The downstream traffic on a wideband interface intermittently stops being forwarded. The queue lengths grows to max as shown in show pxf cpu queue wideband-Cable A/B/C:D command with no further dequeues.
Condition: This issue occurs if downstream bonding group might exist on the interface and more number of channels are present in the bonded channel.
Workaround: There is no workaround, however try reloading microcode to clear the problem if it occurs.
Open Caveats — Cisco IOS Release 12.2(33)SCB1
The open caveats are grouped based on their component types. The open caveats documented from prior releases that are unresolved might also apply to this release.
CMTS-Application
•
Symptoms: Cable metering collection enters "write-error" state and does not recover.
Conditions: When cable metering is configured with default TCP parameters.
Workaround: Enabling "ip tcp path-mtu-discovery" may help prevent occurrences of the issue. Also, the command test cable metering abort will clear the "hung" state and allows the next iteration of cable metering to occur.
•
Symptoms: Throughput rate shows (slightly) more than the expected range.
Conditions: This happens when DMIC feature is enabled and modem is marked with given Max rate values. This can also happen when QoS profile has lower Max rates.
Workaround: There are no workarounds.
•
Symptoms: Certain PCMM calls using wideband CMs may have one way audio.
Conditions: PCMM calls between w-online(pt) and online(pt) on wideband CMs.
Workaround: There are no workarounds.
•
Symptoms: Multicast stream drops after changing ip igmp query-interval
Conditions:
–
–
–
Workaround: There are no workarounds.
•
Symptoms: Inactive sessions are observed in the multicast explicit tracking database.
Conditions: IGMP leave received for a group followed by a cable line card switchover
Workaround: There are no workarounds.
•
Symptoms: SP messages are logged in RP console but not sent as traps by RP's SNMP agent.
Workaround: There are no workarounds.
•
Symptoms: When modem comes up during the off-peak time, when it is not supposed to be monitored as per enforce-rule configured, the smp is not synced to standby line card.
Conditions: If a line card switchover is done during the off-peak time, even if the modem enters peak-time, it is not monitored as per the enforced rule configured. The smp is synced to standby line card only once the modem is monitored during peak-time.
•
Symptoms: Cable metering fails and enters a hung state.
Conditions: This problem happens when the ip tcp timestamp option is configured globally.
Workaround: Do not use the ip tcp timestamp option.
•
Symptoms: Cable modems stuck in expired(pt) state, when both pre-equalization and BPI+ are configured.
Conditions: Certain dynamic faults within the HFC network causing a requirement to adjust the main tap of the modem's transmit equalizer.
Workaround: For MC520s and u cards, prior to 12.3(23)BC6, and CSCsu95526
Use following modulation profile for 64qam, 6.4 MHz channel-width at 2 ticks.
cab modulation-prof 224 request 0 16 0 48 qpsk scram 152 no-diff 32 fixed qpsk0 1 2048
cab modulation-prof 224 initial 5 34 0 48 16qam scram 152 no-diff 384 fixed qpsk1 1 2048
cab modulation-prof 224 station 5 34 0 48 16qam scram 152 no-diff 384 fixed qpsk1 1 2048
cab modulation-prof 224 a-short 6 76 3 22 64qam scram 152 no-diff 64 short qpsk1 1 2048
cab modulation-prof 224 a-long 9 232 0 22 64qam scram 152 no-diff 64 short qpsk1 1 2048
cab modulation-prof 224 a-ugs 9 232 0 22 64qam scram 152 no-diff 64 short qpsk1 1 2048
or if they want more robust VoIP;
cab modulation-prof 224 a-ugs 9 232 0 22 16qam scram 152 no-diff 64 short qpsk1 1 2048
For 16Qams:
cab modulation-prof 223 request 0 16 0 44 qpsk scram 152 no-diff 32 fixed qpsk0 1 2048
cab
CMTS-Database
•
Symptoms: Memory leak at buffer_peak_notify is observed.
Workaround: There are no workarounds.
•
Symptoms: When a UBR10k has a lots of CMes, the command show cable modem docsis version summary total may not be able to display each column properly because there is not enough space between columns.
Conditions: When there are lots of cable modems.
Workaround: There are no workarounds.
•
Symptoms: For deleted service flows the last character in the Service Class Name field is dropped in SAMIS records as well as in the SNMP MIB docsQosServiceFlowLogServiceClassName.
Workaround: There are no workarounds.
CMTS-DOCSIS
•
Symptoms: Some DS packets of UGS traffic is dropped for seven seconds after CM sends another DSC for the service flow.
Conditions: Cable service flow activity-timeout <#> is set to a value shorter than the call duration. For instance, an Arris CM seems to ignore this value, while Scientific Atlanta CM starts resending DSC messages every (period - 10) seconds.
Workaround: Increase cable service flow activity-timeout, or use a different CM model.
•
Symptoms: MDD change message is generated on all the RF channels globally whenever any RF channel is configured on the CMTS. MDD change is supposed to be generated for channels within a MacDomain.
Workaround: There are no workarounds.
•
Symptoms: When UCD change occurs, some data structures related to US scheduler of secondary upstream service flows do not get properly updated. For instance, minislot size. Because of this, any traffic through these service flow might get affected. For instance, packet drop.
Workaround: There are no workarounds.
•
Symptoms: The number of downstream PHS rules for a cable MAC domain stays the same even though more downstreams can be added to the MAC domain (through modular downstreams).
Workaround: There are no workarounds.
•
Symptoms: In a very rare scenario, after an N+1 switchover, upstream traffic that is BPI encrypted is not received properly by the CMTS. Instead, input errors are logged on the interface and "debug cable error" gives error messages similar to:
Cable5/1/4: Bad rx packet. JIB status code 0xA
Conditions: The issue occurs on upstream channels that use a "shared" connector where the other upstream channel using the same shared connector is on another downstream and is shutdown.
Workaround: Activate the downstream and other upstream channel using the same shared connector or unshare the upstream connector temporarily.
CMTS-Management
•
Symptoms: SNMP GetNext requests for docsQosServiceFlowPrimary (also known as 1.3.6.1.2.1.10.127.7.1.3.1.8 or docsQosServiceFlowEntry.8) are rejected. Still, if a certain docsQosServiceFlowPrimary entry is polled with SNMP Get directly (after some additional calculations are performed to determine the index value), the value is returned as expected.
Conditions: UBR7114E running 12.3(21a)BC3.
Workaround: Poll the individual values following the steps of the procedure suggested in SR 610144513.
CMTS-Platform
•
Symptoms: The cable line card crashes after the execution of the commands clear arp and clear cable modem all delete.
Workaround: There are no workarounds.
•
Symptoms: When either Cisco 5x20 card Mac domains have WB associations, a traceback may be observed along with a message similar to the following when NO CARD x/x is used:
*Jan 25 05:41:57.235 PST: %SYS-3-BADLIST_DESTROY: Removed a non-empty list(2F3F7BF4, name: (unknown)), having 36 elements -Process= "C10K Card Actions", ipl= 0, pid= 75A small amount of memory leak may also happen. No other side effects are expected.
Conditions: This traceback is seen only if the card had a NB-WB associations previously, and was powered down or removed.
Workaround: There are no workarounds.
•
Symptoms: Remote modems are using stale blaze indexes not known by Guardian.
Conditions: There are modems online from SPA downstream. Type clear cable modem all reset. Modems come online from SPA downstream again.
Workaround: When you want to reset all the modems, type clear cable modem all delete.
•
Symptoms: In case of multicast and the forwarding is through WB interface, multicast SFs are created and the counters are maintained under the WB interfaces.
Workaround: There are no workarounds.
•
Symptoms: When MVPN is configured on CMTS, and a CPE joins a multicast group that has a source already transmitting to, then the (S,G) entry is not created in mvrf table of the CMTS.
Workaround: There are no workarounds.
•
Symptoms: PRE HA test shows that sometimes the voice call dropout time is larger than one second.
Conditions: Test the offnet call.
Workaround: There are no workarounds.
•
Symptoms: In numerous 10Ks running 12.3(21a)BC7 and BC8, the customer is seeing instances where the CPE device disappears from the cable host table and sometimes from the ARP table. In all cases, the CPE loses IP connectivity. This shows up across different CM vendors, different CLC types, different plants, and different 10Ks.
Conditions:
–
–
–
–
Workaround: Try clear cable host until the device appears back in the host table. This may not work always.
•
Symptoms: Show interface cable x/y/z SID will show null MAC address and null IP address SID entries.
Conditions: IOS 12.3(17b)BC9, UBR10K CMTS
Workaround: There is no workaround.
•
Symptoms: From logs collected from the customer's network, at least three routers show the symptom of the problem. For the two SPAs installed in the Cisco Wideband SIP jacket card, SPA1/0/1 lost all its configurations, including its IP and MAC addresses, and all its RF channel configurations. SPA1/0/0 lost some of its configurations. IP and MAC addresses remain but some of the RF channel configurations are missing. Recovery measures were attempted, including PRE S/O, hw-module reset, unprovision/reprovision the SPAs, but none succeeded. Reloading the router seems to be the only way to recover.
Workaround: Reloading the router is the only way to recover.
•
Symptoms: On a UBR10k the input rate counter of all GRE tunnels is incorrect. It is doubled compared to the real value. Only tunnels interfaces are affected.
Conditions: This occurs on a UBR10k/ESR-PRE2 running IOS versions 12.3(23)BC1 and 12.3(23)BC2 and 12.3(23)BC5.
Workaround: There is no workaround.
•
Symptoms: When upgrading IOS and hardware, the existing configuration could conflict with the channel IDs on other interfaces.
Conditions: Customer would like to see the DS channel IDs hidden, but viewable with a show command, in order to avoid potential conflicts.
Workaround: Reconfigure downstream channel-id for any interface that has a conflicting ID.
•
Symptoms: A Cisco router may reload reporting a bus error.
Workaround: There is no workaround.
•
Symptoms: Backup DTCC card may experience an unexpected reload irregularly with following messages:
Oct 14 23:01:25.487: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:00:00:07 LKA:00:00:07 LRMV:never, NKA:00:00:00, cpu 1%/1% 2%
Oct 14 23:01:25.491: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:never LKA:never LRMV:never, NKA:00:00:00, cpu 1%/1% 2%
Oct 14 23:01:25.491: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is down. Notifying 2cable-dtcc driver.
Oct 14 23:01:25.491: %UBR10K_ALARM-6-INFO: ASSERT CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1
Oct 14 23:01:26.555: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:26.555: %IPCOIR-5-CARD_LOADING: Loading card in slot 1/1
Oct 14 23:01:27.767: %C10K-5-LC_NOTICE: Slot[1/1] Line-card Image Downloaded...Booting... Oct 14 23:01:45.580: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:45.656: %UBR10KTCC-4-CHG_CLK_REF: Clock reference source set to DTI for DTCC card 1/1
Oct 14 23:01:45.936: %UBR10K_ALARM-6-INFO: CLEAR CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1
Oct 14 23:01:45.936: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is up. Notifying 2cable-dtcc driver.
Oct 14 23:01:47.040: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active fast
Oct 14 23:02:20.042: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active normalConditions: This problem is seen on only Backup DTCC irregularly of UBR10K running 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: PRE crashes after removing policy map.
Conditions: This is seen on a UBR10k with PRE3 running 12.3(21a)SB5 IOS.
Workaround: Do not remove policy map.
•
Symptoms: Swapping unlike MC520 line cards (s, u) causes modems go offline, and configuration lost.
Conditions: Default behavior of the cr10k card [slot/subslot]; oir-compatibility command must be converted from default disabled to default enabled for all cable line cards.
Workaround: Prior to exchanging line cards, configure oir-compatibility for all slots. If linecard exchange happens without configuring oir-compatability, and the problem is discovered before a wr mem is issued, perform the following workaround:
copy sec-nvram:startup-config to external box
edit card types from MC520s-d to MC520u-d
copy modified file to nvram:startup-config, and also sec-nvram:startup-config
reload
This is the only procedure which can ensure that the frequency stacking and virtual interface configurations are preserved. Attempts to paste pieces of the previously stored running config will fail if frequency stacking or virtual interfaces are configured, because the connectors must be un-assigned first.
If a wr-mem has occurred, then the shutdown state and blank config for all interfaces will be written to both primary and secondary nvram.This technique will not work without resorting to an externally stored backup configuration for the system.
CMTS-Redundancy
•
Symptoms: Modems on modular interface suffer long-time recovery time or even fall offline after switchover or revertback.
Conditions: Only affect 520S/520U card as active card. Only affect modems on modular interface. More visible for annex A modulation type.
Workaround: There is no workaround.
•
Symptoms: Memory leak at cmts_hccp_get_tlv_parser_dat_gbl has been observed.
Workaround: There is no workaround.
•
Symptoms: Both the working cards 5/0 and protect card 5/1 are active.
Conditions: Do a LC SO from 5/0 to 5/1. After the synced, shut down 5/0/2 interface. Do a LC revertback from 5/1 to 5/0. Do no shut 5/0/2 (Do not wait in between step3 and step 4). Now you will find, after suspend timer expires, both 5/0 and 5/1 are active.
Workaround: Do not shut down an active protect interface. If both the cards (working and protect) are active, reset protect card.
•
Symptoms: During LCHA, the CM instance bpi_enable flag (cminstp->priv_en) on RP is not synced over correctly. This is observed for MC interface CM with bighorn image.
Conditions: Generate multicast stream to CMs, bpi CM during LC switchover.
Workaround: There is no workaround.
•
Symptoms: Secondary PRE crashes when DBS function is applied to wideband interface.
Conditions: Problem is seen on a PRE-2 running 12.3(23)BC2. To apply DBS, the customer has to shut down the wideband interface first and then they met a IOS crash when "chunk_free(docsis_srv_flow_chunk,flowp)" at the last step of cmts_destroy_ds_service_flow on secondary PRE. No PXF happens.
Workaround: There is no workaround.
CMTS-Voice
•
Symptoms: 12.3(23)BC2 mixed docsis mode (atdma-tdma) can only achieve 40 UGS for an upstream. But, 12.3(17b)BC9 can achieve 48 UGS flows.
Conditions: Currently, it looks like this may have something to do with mixed mode and modems.
Workaround: There is no workaround.
•
Symptoms: When TWC, the serviceflow in admitted state still has traffic going through.
Conditions: In three-way calling, and one side is on hold.
Workaround: There is no workaround.
CMTS-Wideband
•
Symptoms: The show controller modular-cable 1/0/0: command shows that rf-channel 0 become unknown after hw-module reset.
SPA IP address = 192.168.5.1 SPA MAC Addr = 0012.001A.67D7 QAM MOD ANNEX TKB Interval Rate adjust State 0 Unknown Unknown 0 1 Enabled ====> unknown 1 QAM 256 Annex B 2423 132 Enabled 2 QAM 256 Annex B 2423 132 Enabled 3 QAM 256 Annex B 2423 132 Enabled
Conditions: Happened after hw-module reset and show diag 1/0/0. This will cause all wideband modems also go to NB.
Workaround: Reload the router to correct the issue.
Resolved Caveats — Cisco IOS Release 12.2(33)SCB1
The resolved caveats are grouped based on their component types.
CMTS-Application
•
Multiple Cisco products are affected by denial of service (DoS) vulnerabilities that manipulate the state of Transmission Control Protocol (TCP) connections. By manipulating the state of a TCP connection, an attacker could force the TCP connection to remain in a long-lived state, possibly indefinitely. If enough TCP connections are forced into a long-lived or indefinite state, resources on a system under attack may be consumed, preventing new TCP connections from being accepted. In some cases, a system reboot may be necessary to recover normal system operation. To exploit these vulnerabilities, an attacker must be able to complete a TCP three-way handshake with a vulnerable system.
In addition to these vulnerabilities, Cisco Nexus 5000 devices contain a TCP DoS vulnerability that may result in a system crash. This additional vulnerability was found as a result of testing the TCP state manipulation vulnerabilities.
Cisco has released free software updates for download from the Cisco website that address these vulnerabilities. Workarounds that mitigate these vulnerabilities are available.
This advisory is posted at http://www.cisco.com/warp/public/707/cisco-sa-20090908-tcp24.shtml.
•
The server side of the Secure Copy (SCP) implementation in Cisco IOS software contains a vulnerability that could allow authenticated users with an attached command-line interface (CLI) view to transfer files to and from a Cisco IOS device that is configured to be an SCP server, regardless of what users are authorized to do, per the CLI view configuration. This vulnerability could allow valid users to retrieve or write to any file on the device's file system, including the device's saved configuration and Cisco IOS image files, even if the CLI view attached to the user does not allow it. This configuration file may include passwords or other sensitive information.
The Cisco IOS SCP server is an optional service that is disabled by default. CLI views are a fundamental component of the Cisco IOS Role-Based CLI Access feature, which is also disabled by default. Devices that are not specifically configured to enable the Cisco IOS SCP server, or that are configured to use it but do not use role-based CLI access, are not affected by this vulnerability.
This vulnerability does not apply to the Cisco IOS SCP client feature.
Cisco has released free software updates that address this vulnerability.
Workaround: There are no workarounds available for this vulnerability apart from disabling either the SCP server or the CLI view feature if these services are not required by administrators.
This advisory is posted at the following link:
http://www.cisco.com/warp/public/707/cisco-sa-20090325-scp.shtml.
•
Cisco IOS software contains a vulnerability in multiple features that could allow an attacker to cause a denial of service (DoS) condition on the affected device. A sequence of specially crafted TCP packets can cause the vulnerable device to reload.
Cisco has released free software updates that address this vulnerability.
Several mitigation strategies are outlined in the workarounds section of this advisory.
This advisory is posted at http://www.cisco.com/warp/public/707/cisco-sa-20090325-tcp.shtml
•
Symptoms: Modems may go offline with tdma-atdma mode if the US is configured with minislot size as 4.
Workaround: Change the mini-slot size to 2.
•
Symptoms: While configuring a modular cable interface on cable SPA 3/3/0, there is a memory corruption on the mac-domain host line card. This eventually causes the mac-domain host line card to crash. If line card HA is configured, then after the line card crashes, an HA switchover to the Protect LC is triggered, and the Protect LC will now start to crash.
Conditions: Cable SPA is in slot 3, bay 3, and default multicast QoS is configured on the CMTS.
Workaround: 1) Do not use slot 3 and bay 3 for cable SPA or 2) Do not configure multicast QoS on the CMTS (including multicast QoS related to DSG).
•
Symptoms: Memory corruption on cable line card when running CMGEN image and CM GEN tests.
Conditions: CMGEN tests.
Workaround: There is no workaround.
•
Symptoms: When a Cisco uBR series CMTS is configured to perform CPE IP to MAC address resolution through the use of DHCP LEASEQUERY messages rather than ARP, the address resolution will fail when it is inspired by downstream packets to the unresolved destination.
Workaround: Temporarily allow downstream ARP resolution with the cable bundle interface commands cable arp and cable proxy-arp.
•
Symptoms: Show interface x/y/z modem 0 may show duplicate MAC entries.
Workaround: There is no workaround.
•
Symptoms: MC520U card crashes due to TLB BUS error when too many "cable modem xxxx change-frequency yyyy" commands are executed.
Conditions: More than one DS ports are connected to one HFC. When one line card crashes, modems can be onlined to other line card's DS port. LB is used.
Workaround: There is no workaround.
•
Symptoms: DBC w/ DS physical change invoked through the VDOC broadcast feature fails on the SA-3000.
Conditions: VDOC broadcast RF-span feature is used w/ SA-3000 CMs.
Workaround: There is no workaround.
•
Symptoms: Multicast session of static IGMP group mess up after the LC switchover.
Conditions: Config static messiness group on if and do LC S/O.
Workaround: Do not config static session or do not let multicast client rejoin after a while.
•
Symptoms: IGMP query messages are not forwarded to the multicast client from CMTS when modems are registered in modular interface with TLS. Due to this, packets are not received at the client end.
Conditions: TLS and Multicast.
Workaround: There is no issue with legacy setup.
•
Symptoms: TLS data traffic either stops or forwarded uninterrupted after OIR of cable line card.
Conditions: TLS is provisioned on a CM with modular cable primary downstream and the modular host is on a different line card than the MAC-domain host interface of the CM.
Workaround: There is no workaround.
•
Symptoms: Removing the dot1q vc map causes WAN interface to flap.
Conditions: uBR10k chassis running 12.3BC release. Removing the last dot1q vc map in the running configuration caused the WAN interface to flap.
Workaround: There is no workaround.
•
Symptoms: Multicast traffic over DOCSIS L2VPN and TLS on secondary RF-channels of a D3.0 CM is dropped by the CM.
Conditions: CM is MDF disabled.
Workaround: There is no workaround.
•
Symptoms: The CMIM mask for pre-reg DSID is not correct. It should include the CM stack as well.
Conditions: The SA in REG-RSP/REG-RSP-MP should be in Dyanmic type for TLV64 request.
Workaround: There is no need to include GMAC for static DSID for MDF2 CM.
•
Symptoms: Break compiling cr7200 image.
Conditions: Compiling cr7200 image.
Workaround: After this fix, cr7200 image can be compiled.
•
Symptoms: After using dcc technique 1~4 move a modem from one channel to another channel, if this modem has a rsvp flow then this rsvp flow db in rp will be lost after dcc.
Conditions: If modem has a rsvp service flow, load balance is enabled, and dcc-init-technique of load balance group is from 1 to 4.
Workaround: Configuring dcc-init-technique of load balance group to 0.
•
Symptoms: Stale dynamicSF shown after PRE switchover.
Conditions: Moving CM which has dynamic service-flow created (dynamic service-flow created through VDOC RSVP bandwidth reservation) to different upstream, and during PRE switchover delete and re-added RSVP reservation.
Workaround: There is no workaround.
•
Symptoms: Downstrean dynamic service flow reserve CIR release from wrong channel if it change from one dowstream channel to another because attribute mask change when DSC.
Conditions: Create a dynamic service flow with reserve CIR in one downstream channel with attribute-mask configured. Change this attribute configuration, then if dynamic service flow is changed, it maybe reassigned to another downstream channel, then the reserve bandwith would be released from new assigned channel but not original channel.
Workaround: Do not change attribute-mask configuration when there are downstream dynamic service flows created in it.
•
Symptoms: After PRE switchover, the throughput of MC/WB displayed by running the command show interface X/Y/Z is much larger than the real value. After 10-20 minutes, the throughput recovers automatically. On 12.3(23)BC, this also happens on legacy interface.
Conditions: PRE switchover.
Workaround: There is no workaround.
•
Symptoms: When TACACS is configured, no aaa new-model authentication happens through local. This happens for the exec users under vty configuration.
Conditions: Configure no aaa new-model
Configure login local under line vty 0 4
Configure login tacacs under line vty 0 4Workaround: There is no workaround.
•
Symptoms: On a router with CEF switching turned on, a memory leak is observed for ARP packets.
Workaround: There is no workaround.
•
Symptoms: Ping to a /31 loopback interface in VRF fails.
Conditions: A loopback with a /31 address configured in a VRF.
Workaround: There is no workaround.
•
Symptoms: Three separate Cisco IOS hypertext transfer protocol (HTTP) cross-site scripting (XSS) vulnerabilities and a cross-site request forgery (CSRF) vulnerability have been reported to Cisco by three independent researchers.
Workaround: Disable the HTTP server.
•
Symptoms: When the routers are booted sequentially (bringing up the standby after the active boots up), the routers fall into RPR instead of SSO.
Conditions: Seen with NAM module with SRC2 image.
Workaround: There is no workaround.
•
Symptoms: Improve debugability of CMTS chassis
Workaround: show tech-support
show tech-support cmts c\x\y\z
show tech-support cmts modem <ip/mac>
•
Symptoms: Cable trust does not seem to have any impact on source verification.
Conditions: Cable trust command can be used to prevent the CMTS from sending out DHCP LQ when the source verification fails for a given CPE MAC address. But in spite of such a configuration, CMTS seems to be sending out DHCP LQ when the source verification fails.
Workaround: There are no workarounds.
•
Symptoms: CMIpAddr may return 0 after a line card switchover.
Conditions: Configure ipdr and collector.
Workaround: There are no workarounds.
•
Symptoms: WB interface is in link state down after a PRE switchover. This causes the static mcast forwarding to stop for the WB interface and causes static mcast to be created dynamically and forwarded.
Conditions: The WB interface only goes into this error condition if the particular interface is brought link up after the standby-pre is already up, and PRE switch over occurs.
Workaround: Do not bring a WB interface up after the standby-PRE is already up. The other workaround is to do another PRE-switchover to go back to a new active PRE.
•
Symptoms: Replacing an MC520S/U/H with another MC520S/U/H causes modems to drop offline and generate a console error message similar to the following:
ERROR: 1/0/0 rf-channel 0: already hosted under Ca5/1/1
The message indicates that an attempt was made to add the rf-channel on the Modular-Cable controller in slot 1, subslot 0, bay 0 to another the MAC domain when it was already configured for MAC Domain 1 on the Cable line card in slot 5, subslot 1.
Conditions: The problem occurs when the OIR-compatibility feature is invoked to preserve a line card configuration across an OIR operation when the line card configuration contains remote Modular-Cable downstreams and N+1. The compatibility feature is only invoked when a change of card type is detected. Replacing an MC520 with the same card type will not cause the problem.
Workaround: There are no workarounds.
•
Symptoms: WB modems reset after replacing an MC520 guardian card with a different type of MC520, that is, an MC520U is replaced by an MC520H.
Conditions: Wideband modems will flap in an N+1 configuration under the following conditions:
1. The Guardian is included in a line card N+1 redundancy configuration
2. OIR compatibility is enabled on the working guardian's slot/subslot
3. The working guardian has failed to the protect card.
4. The working guardian is replaced by a different type of MC520.
Workaround: There are no workarounds.
•
Symptoms: WB modems resets after LC SO of guardian card through OIR method.
Workaround: Switchover the working card to the protect card before OIR the working card.
•
Symptoms: Aggregate MQoS session statistics not updated correctly.
Conditions: This issue occurs when aggregate MQoS is configured.
Workaround: Single MQoS session could be configured.
•
Symptoms: IPDR DOCID is incorrect. The first 8 hex digits of the docid are supposed to be seconds since the Unix epoch. The collector shows:
0x48efa0fa = 2008-10-10 18:37:46.00
The CMTS reports:
0xCC98B503 = 2077-11-22 19:27:56.00
Conditions: This issue occurs when aggregate MQoS is configured.
Workaround: The current value subtract 70 years in seconds from 1 Jan 1900 to 1 Jan 1970.
CMTS-DOCSIS
•
Symptoms: When users use virtual interface, it is possible that Ingress Noise Cancellation functionality on the upstream is wrongly activated or deactivated.
Conditions: 1) User used odd logical upstream channel ID with even physical port ID or even logical upstream channel ID with odd physical port ID. 2) Only affect 5x20H/20x20H/2x8U cards.
Workaround: Use odd logical upstream channel ID with odd physical port ID and even logical upstream channel ID with even physical port ID.
•
Symptoms: show cable modem wideband shows all MD-DS-SG as 1, and show cable mac-domain rcc give different values for the MD-DS-SG that modems are assigned to.
Workaround: There is no workaround.
•
Symptoms: When an RCC template is configured and applied to an interface(for example a 4 channel RCC), and a fiber node is added containing only 3 out of the 4 channels, the CMTS incorrectly creates an RCC association with the MD-DS-SG for that fiber node with 3 channels.
Workaround: Remove and reapply the RCC template from the interface.
•
Symptoms: Channels with center frequency greater than 850 MHz cannot be inputted as receive-channels in rcc-templates. The SPA allows max frequency of 1050 MHz channels.
Workaround: Do not use RCC-templates when needing to use higher frequency than currently supported 850.
•
Symptoms: Certain cable modems are shown stuck in the sreject(na) state on rare occasions.
Conditions: In the event a cable modem returns a confirmation-code of partial-service (30) in the REG-ACK message, the CMTS treats this as an error, thus leaving the CM in the sreject(na) state.
Workaround: There is no workaround.
•
Symptoms: CMs stuck in init(i), init(o) after a TFTP server outage with Dynamic Shared Secret on interfaces configured with DMIC Reject or Marked.
Conditions: Seen only on the interfaces which are configured with DSS. Seen only when the TFTP server has some reachability problems. Seen in both ubr7200 and ubr10k and different IOSes (12.3(9a)BC6, 12.3.13a-BC1,12.3(23)BC2)
Workaround: Remove DSS from the cable interface.
•
Symptoms: Using the test cable dcc command from the modular to the native interface moves the CM just fine. Using the test cable dcc command from the native to the modular drives the CM offline, at which point it's reported [offline] on another native DS where it can never possibly register by the cable plant design. Finally the CM reaches the right online on the right DS again.
Conditions: Wideband Cable deployment. The CM must fail to move from the native to the modular DS.
Workaround: When the CM is allowed to recover it is reported under a correct interface.
•
Symptoms: CMs stuck in init(i), init(o) after a TFTP server outage with Dynamic Shared Secret on interfaces configured with DMIC Reject or Marked.
Conditions: Seen only on the interfaces which are configured with DSS. Seen only when the TFTP server has some reachability problems. Seen in both ubr7200 and ubr10k and different IOSes (12.3(9a)BC6, 12.3.13a-BC1,12.3(23)BC2)
Workaround: Remove DSS from the cable interface.
•
Symptoms: After wideband cable modem doing DCC, show cable modem info as command "scm", the warning message "Warning: Zero Blaze index for SFID" is displayed.
Conditions: 1. Wideband cable modem on line. 2. Doing DCC.3. show cable modem information as CLI command "scm".
Workaround: There is no workaround.
•
Symptoms: PRE2 crashes when no card xxxx is executed.
Conditions: no utility card, then no any spa on the same slot number.
Workaround: Do not do no utility card.
•
Symptoms: US bitmap is not updated after changing max-ports.
Conditions: Change the upstream max-ports using the command cable upstream max-ports and configure US connector to the new upstream channel.
Workaround: Remove and reconfigure the cgd configuration under that mac-domain if max-port is changed.
•
Symptoms: When a 520H is in normal operation condition, modems on one or a few upstreams get into a bad state, and all modems on the affected upstreams go offline, while other upstreams are still functioning. This affects only 520H LC on all releases.
Conditions: Traffic on upstreams.
Workaround: Issue a shut/no shut on the affected upstreams recovers it, and modems come back online. Upgrade to 12.3(23)BC6.
•
Symptoms: Docsis 2.0 modems stuck in reject(na) on a CMTS with 12.2(33)SCB code.
Conditions: Registering a modem using service class names for service flow definitions.
Workaround: Do not use service class names in modem configuration, or make sure that modem firmware is such that unknown service flow TLVs are ignored instead of rejected.
CMTS-Drivers
•
Symptoms: Ironbus restarts have been observed on the Cisco uBR10012 due to ironbus link status 0x1180 errors. You observe the following messages in the PRE log: ------------------------------------------------------------------------------------------------------ 811468: Jun 17 03:10:17.233 UTC: slotindex is 8. 811469: Jun 17 03:10:17.233 UTC: IB Link status: 00001180 811470: Jun 17 03:10:17.233 UTC: %C10KEVENTMGR-1-IRONBUS_FAULT: Ironbus Event 5/0, Restarting Ironbus 811471: Jun 17 03:10:17.645 UTC: %C10KEVENTMGR-1-IRONBUS_SUCCESS: Ironbus Event 5/0, Restart Successful -----------------------------------------------------------------------------------------------------
The ironbus link status 0x1180 error will trigger an line card switchover on uBR10012s configured with N 1 redundancy. The ironbus restart is fast enough to keep modems online and has negligible affect to customers on uBR10012s without N 1 redundancy.
Conditions: This ironbus link status 0x1180 error has only been observed on slot 5/0 with the following hardware configuration. - Working MC520H Linecard in slot 5/0 (the problem has not been reported on the MC520U/S cards) - Active PRE2 in slot A (the problem has not been reported on a PRE1) The CMTS BU is still investigating the root cause of the ironbus link status 0x1180 error.
Workaround: There are two workarounds for ironbus link status 0x1180 error 1. Use Active PRE2 in slot B 2. Do not use a MC520H in slot 5/0 Customers should not RMA any equipment due to the ironbus link status 0x1180 error. The CMTS BU has not confirmed that this is a hardware issues and would need to identify the faulty hardware (e.g., PRE2, MC520H, chassis) if it is a hardware issue.
CMTS-ISSU
•
Symptoms: When a CLC is Do not running at 99% CPU load, IPC timeout triggered traceback will be seen on the CLC, or CLCs that sending IPC messages to this LC.
Conditions: When a CLC has high CPU usage.
Workaround: Try to find out the cause of high CPU usage. If it is due to excessive SNMP polling, reduce the SNMP polling.
•
Symptoms: IPC connection between RP and CLC may take one extra second to come up at ISSU runversion.
Conditions: At issu runversion, when standby RP becomes active with the new version.
Workaround: There is no workaround.
•
Symptoms: The purpose of this enhancement is to restrict access to the HA redundancy cmds during the "issu linecard changeversion" process.
The following additional commands are restricted during uBR10K ISSU. This prevents the RF switch configuration from being altered during upgrade, and avoids switchover failures because of that.
Router(config-red-lc)#rf-switch snmp-community xxx
Router(config-red-lc)#rf-switch name xxx
Router(config-red-lc)#rf-switch protection-mode xxx
•
Symptoms: Double commit fix from cobra_cable to throttle rebuild 1.
Workaround: There is no workaround.
CMTS-Management
•
Symptoms: Bad dequeue error is seen at "CMTS METERING EXPORT Process".
Conditions: When 45K cable modems and IPDR are configured.
Workaround: There is no workaround.
•
Symptoms: ccwbFiberNodeNBIfIndx returns 0 if item's rfid belongs to modular interface.
Conditions: Configure fiber node with donwstream modular cable and then getmany -v2c ccwbFiberNodeNBIfIndx
Workaround: There is no workaround.
•
Symptoms: IPDR data collection from CLC to RP might congest the BPE due to its high bursty traffic pattern.
Workaround: This fix aims at throttling the data collection and added control to the throttling via a new command.
command addition: (config) [no] cable metering data-per-session X timer Y
Range X: 3-30
Range Y: 8ms - 500ms
Default X: 5 (size of each data is approx 1KB. So 5KB burst)
Default Y: 100ms
The above command controls the number of IPDR data sent per session (X) in an interval of time (Y). It will be visible only if one changes the default settings. It cannot be changed during data export. The show cable metering verbose command is extended to show the data-per-session and timer when the IPDR data export is IN Progress.
In addition, the range of the command has been changed. Cable metering destination range has been reduced to 2 minutes. 2min-1440min
Due to the newly added IPDR throttling mechanism the time to export of all the IPDR data is longer. It would approximately be time_to_export (minutes) = ( total_flows * Y) / ( X * 1000 * 60 ), where X and Y are configurable via the newly added command.
CMTS-Platform
•
Symptoms: PRE2 crashed multiple times at sch_rp_cm_reg_info_request_handler process.
Workaround: There is no workaround.
•
Symptoms: The bug causes ironbus restart on a reload.
Conditions: When reload is typed on RP console, sometimes user could see "Ironbus restarting" prints on RP console.
Workaround: There is no workaround.
•
Symptoms: When the following commands are run from RP or LC continuously, it causes line card CPUHOG, IPC timeout and line card crash or reset.:
test cable phydump c7/1/0
test cable phydump c7/1/1Workaround: Do not run "test cable phydump c7/1/0" and add some sleep between commands. Run the command like:
1. test cable phydump c7/1/0 phy 0
2. tes cable phydump c7/1/0 phy 1
3. sleep 1s
1. test cable phydump c7/1/0 phy 2
2. tes cable phydump c7/1/0 phy 3
•
Symptoms: Modular modems offline when changev all minor or major MDR is executed.
Conditions: The ISSU pre runversion will cause sequentially reload of both DTI cards. After the reload, only one card is up and the port of the other card is down with no signal.
Workaround: There is no workaround.
•
Symptoms: sinit and sreject states show up on a 10k running Amazon (12.2(33)SCB). However, no EAE is configured.
Conditions: Doing an HCCP failover.
Workaround: There is no workaround.
•
Symptoms: Modems go off line.
Conditions: pre-eq is enabled with ATDMA 64QAM 6.4Mz.
Workaround: Use a different modulation profile.
•
Symptoms: The Cisco MC5X20U cards may experience a silent reload and no crashinfo file is written to bootflash.
Conditions: This is seen on a PRE 2 processor moule running Cisco IOS Release 12.3(13a)BC2.
Workaround: Upgrade to Cisco IOS Release 12.2(33)SCB2.
•
Symptoms: RPC message to read EEPROM of SPA is failed because RPC timeout.
Conditions: When SPA CPU is occupied by a task for a long time, for instance FPD reprogramming task is updating FPGAs.
Workaround: There is no workaround.
•
Symptoms: show diag 2/1 for TCC not available on CMTS.
Conditions: Using 122SC based IOS image.
Workaround: There is no workaround.
•
Symptoms: Blaze index is not cleared from service flow when WB interface is shut.
Conditions: When WCM is on w-online. After Wide-Band interface is shut, WCM is offline, and when it comes back with modular online, it uses the original Blaze Index.
Workaround: There is no workaround.
•
Symptoms: A new message appears in the console log indicating that a card type has changed. For example, if the MC520U-D in subslot 8/1 is replaced by an MC520S-D, the following message is displayed:
%UBR10K-6-COMPAT_NEW_CARD: The 5cable-mc520u-d in slot 8/1 has been replaced by a 5cable-mc520s-dConditions: The message appears when an OIR operation involves two different types of MC520 line cards. For example, when an MC520U is replaced by an MC520H. The message does not appear for cards other than the MC520S/U/H. It also does not appear if the OIR operation involves identical MC520 card types. For instance, it will not appear if an MC520U is replaced by another MC520U.
Workaround: This is not a failure so no workaround is needed.
•
Symptoms: When setting the Fragment-force 2000 the upstream performance drops to zero. There is no upstream throughput.
Conditions: Problem is found only on the 5x20s card with ios ubr10k2-k8p6u2-mz.123-23.BC1.
Workaround: Put Fragment-force to Frag-force 1987 for best performance.
•
Symptoms: Incorrect statistics displayed due to loss of periodic push statistics packets.
Conditions: Problem is found only on the 5x20s card with ios ubr10k2-k8p6u2-mz.123-23.BC1.
Workaround: There is no workaround.
•
Symptoms: On a 10k running 12.2(33)SCB, this bug is to address an issue where the controller Modular-Cable will show up in the wrong order in running-config if more than one is inserted into the chassis.
Conditions: More than one 24XDS SPA inserted into a chassis where 12.2(33)SCB is running.
Workaround: There are no workarounds.
•
Symptoms: Since the sflow statistics are pushed to the 5x20 cable line card every ten seconds, there is a potential to lose up to ten seconds worth of statistics since this IPC message will reset the blaze hw.
Conditions: This could happen every ten seconds.
Workaround: There are no workarounds.
•
Symptoms: 12.2(33) beta and FCS images create default channel IDs that are duplicated across slots 5/0 through 8/1 on the uBR10k chassis.
Conditions: Configuring the same downstream channel from two different linecards to the same fibernode will create an invalid fibernode due to the channel id conflict and therefore requires the user to reconfigure one of the channel ids to a value other than the default. Fibernode status can be seen with show cable fibernode.
Workaround: Manually reconfigure the downstream channel id such that they do not conflict in the fibernode.
•
Symptoms: 520H card crash.
Conditions: PREs s/o. Because it is a timing issue, it does not happen every time. In fact, it is hard to reproduce.
Workaround: There is no workaround.
CMTS-PRE4
•
Symptoms: Memory leak is seen after running the command show debug debug leaks.
Workaround: There is no workaround.
CMTS-PXF
•
Symptoms: When applying a QoS policy to the main interface on a CMTS uBR10k router, the policy does not work for the subinterfaces.
Conditions: QoS applied on an interface that has subinterfaces configured.
Workaround: Configure the service-policy on the main interface before creating the subinterface.
•
Symptoms: This problem is seen when doing source-verify on multicast packets. If the source IP address requires a final-level lookup in the M-TRIE, source-verify will likely fail. The packet could be diverted or dropped.
Conditions: This affects PRE-2 and PRE-4 in 12.2 releases. The 12.3 releases are not affected.
Workaround: There is no workaround.
CMTS-Redundancy
•
Symptoms: Modems are in w-online state with two channel bonding, primary interface on local 520 card. After LCSO, modems still in w-online state and accessible, then clear all modems and wait till online. Some modems come up online in legacy mode instead of wideband mode.
Conditions: Include x/y/0 - x/y/4 in one fiber node. After LCSO, only x/y/0 and x/y/1 could work. MDD for x/y/2 x/y/3 x/y/4 are wrong.
Workaround: Only include x/y/0 and x/y/1 in fiber node.
•
Symptoms:PRE2 crashes when trying to remove rfsw-community from linecard-group.
Conditions: Remove line card from redundancy group.
Workaround: There is no workaround.
•
Symptoms: Wbcm cannot online after removing module-host. From the log, the LC continues to send the MDD to modems, which will push the modems to w-online and cmts denies it.
Conditions: 1. Remove module-host 7/0 2. Adding module-host 7/0 3. LCSO 7/0 4. Remove module-host 7/0.
Workaround: There is no workaround.
•
Symptoms: Modem does not response to ping after PRESO when LCSO occurs in booting Sec PRE.
Conditions: There are two ways to reproduce:
Way 1:
Step:
1 bootup active PRE
2 bootup secondary PRE
3 LC switchover
4 PRE switchover
5 PRE switchover
Way 2:
1 bootup active PRE, LC switchover
2 bootup secondary PRE
3 PRE switchover
Workaround: 1. Before PRE swithover, do LC revert back and then swithover. 2. Only 1 protect 1.
•
Symptoms: WB modem can get wb-online.
Conditions: "shutdown" and "no shutdown" an active protect interface.
Workaround: Do not shutdown an active protect interface.
•
Symptoms: A few cable modems may become offline during the ISSU upgrade, specifically during the ISSU line card prepareversion phase. Those few modems come back online shortly afterwards.
Conditions: There are multiple cable line cards being upgraded through ISSU in sequence, with one or more cable line cards being the modular host. The cable line cards being upgraded subsequent to the modular host cable line cards may encounter this symptom.
Workaround: There is no workaround.
•
Symptoms: Secondary PRE keeps crashing by bus error.
Conditions: After changing the bonding-group-id on wideband interface. After PRE switchover, or hw-module sec-cpu reset.
Workaround: Chassis power off/on, or just reload active PRE.
CMTS-Spectrum
•
Symptoms: Customer is using a customized modulation profile that was working in 12.3(23)BC but is not working in 12.2S. It is causing modems to fall offline and register on a different upstream.
Workaround: 1) Reduce the preamble length of the station ranging burst
cable modulation-profile 123 initial 5 34 0 48 qpsk scrambler 152 no-diff 256 fixed
cable modulation-profile 123 station 5 34 0 48 16qam scrambler 152 no-diff 256 fixed2) set both initial and station to 16qam.
cable modulation-profile 123 initial 5 34 0 48 16qam scrambler 152 no-diff 392 fixed
cable modulation-profile 123 station 5 34 0 48 16qam scrambler 152 no-diff 392 fixedCMTS-Voice
•
Symptoms: Intermittent packet cable call drops on ubr10K. Second attempt is always successful.
Conditions: Running IOS 12.3(17a)BC7.
Workaround: There is no workaround.
CMTS-Wideband
•
Symptoms: When the IP set-top-box (STB) or CPE sends an IGMPv2 leave for a multicast group that is configured for static forwarding based on "ip igmp static-group" interface configuration on the CMTS, then the CMTS stops forwarding the stream.
Conditions: This issue happens when there is a static-group configuration for SSM on the CMTS and CPE is sending ASM join to that SSM group range. When IGMP leave for that SSM range is received, CMTS mistakenly removes the forwarding entries and multicast forwarding stops.
Workaround: Add an input ACL to the bundle interface to block all igmp packets on the upstream for streams that are configured for static forwarding on the CMTS.
•
Symptoms: When using modem-method load balancing, the docsis 3.0 modems are counted as single modem PER downstream channel. This results in many D3.0 modems on 1 channel, which means D2.0 modems will get pushed off, leaving that channel only for bonded traffic (in the case that Dynamic Bandwidth Sharing is in use, with the SPA channel given 90% remaining ratio for both types of traffic i.e. best effort). If the D3.0 modem was counted as a single modem on the Bonded Group of channels i.e. on 3 channels, then they would not influence the movement of D2.0 modems as the D3.0 modems would appear to be equally spread across the SPA channels.
Conditions: D3.0 Modem with 12.3(23)BC5. Load Balance is enabled.
Workaround: There is no workaround.
•
Symptoms: Traceback pops out.
Conditions: 1) Having CGD configuration 2) LC SO 3) LC revertback.
Workaround: There are no workarounds. It does not impact the service.
Open Caveats — Cisco IOS Release 12.2(33)SCB
The open caveats are grouped based on their component types. The open caveats documented from prior releases that are unresolved might also apply to this release.
CMTS-Application
•
Symptom: MC520U card crashes due to TLB BUS error when too many "cable modem xxxx change-frequency yyyy" command are executed.
Conditions: More than one DS ports are connected to one HFC. When one LC crashes, modems can be onlined to other LC's DS port. LB is used.
Workaround: There are no workarounds.
•
Symptom: Periodic 1 second delays between ubr10k transmission of SAMIS packets; enough of these are occurring to make a significant impact on overall SAMIS transfer time.
•
Symptom: Inactive sessions are observed in the multicast explicit tracking database.
Conditions:IGMP leave received for a group followed by a CLC switchover
Workaround:None
•
Symptom: When MVPN is configured on the CMTS, and a CPE joins a multicast group that a source.
Conditions: The issue only happens when igmp join arrives at the same time no mutlicast data sendingis already transmitting to, then the (S,G) entry is not created in mvrf table of the CMTS.
Workaround: There are no workarounds.
•
Symptom: Traceback observed after a uBR10000 PRE switchover %COMMON_FIB-4-FIBNULLHWIDB: Missing hwidb for fibhwidb Tunnel1 (ifindex 144)
Conditions: Multicast VPNs are configured
Workaround: There are no workarounds.
•
Symptom: Throughput rate shows (slightly) more than than the expected range.
Conditions: Happens when DMIC feature enabled and modem marked with given MAX rate values. This can also happen when QoS profile has got lower Max rates.
Workaround: No known workaroud exists. This behaviour is mostly due to smaller value of the MAX-rate, which is not very much applicable with customer scenario. Also the slight change on the throughput values does not have impact on other functionalities
•
Symptom:Downstrean dynamic service flow reserve CIR release from wrong channel if it change from one dowstream channel to another because attribute mask change when DSC.
Conditions: Create a dynamic service flow with reserve CIR in one downstream channel with attribute-mask configured. Change this attribute configuration, then if dynamic service flow is changed, it maybe reassigned to another downstream channel, then the reserve bandwith would be released from new assigned channel but not original channel.
Workaround: Do not change attribute-mask configuration when there are downstream dynamic service flows created in it.
•
Symptom: ToS value on MCAST packets are being incorrectly overwritten.
Conditions: This is in 12.2(33)SCB.
Workaround: There is no workaround
•
Symptom: IGMP join from a cable CPE is timed out improperly by the CMTS.
Conditions: IGMP query interval is set to a large value on the cable bundle interface.
Workaround: There is no workaround.
•
Symptom: When modem comes up during off-peak time when it is not supposed to be monitored as per enforce-rule configured, the smp is not synced to standby LC.
Conditions: If an LC- Switchover is done during off-peak time, Even if the modem enters peak-time, its not monitored as per the enforced rule configured.The smp is synced to standby LC only once the modem is monitored during peak-time.
CMTS-Database
•
Symptom: For deleted service flows the last character in the Service Class Name field is dropped in SAMIS records as well as in the SNMP MIB docsQosServiceFlowLogServiceClassName
CMTS-DOCSIS
•
Symptom: The Cisco uBR10012 doesn't properly reinitialize the sequence number field in the DS extended header of the first packet transmitted on a newly created resequencing DSID. There's no problem in case CM is prepared to receive a non-zero initial sequence number but Cisco uBR should meet the following specification. --- CM-SP-MULPIv3.0-I08-080522 --- 8.2.3 Sequenced Downstream Packets (snip) The CMTS MUST use a Packet Sequence Number of zero (0) in the DS EHDR of the first packet transmitted on a newly created Resequencing DSID.
Workaround: None
•
Symptom: Packet classifier match count is a 32-bit signed counter. It goes negative when it goes above 2 billion packets.
Workaround: There is no workaround.
•
Symptom: The Number of Downstream PHS rules for a Cable MAC Domain stays the same even though more downstreams can be added to the MAC Domain (via Modular Downstreams).
Workaround: There is no workaround.
•
Symptom: Using the test cable dcc command from the modular to the native interface moves the CM just fine. Using the test cable dcc command from the native to the modular drives the CM offline, at which point it's reported [offline] on another native DS where it can never possibly register by the cable plant design. Finally the CM reaches the right online on the right DS again. Conditions: Wideband Cable deployment. The CM must fail to move from the native to the modular DS.
Workaround: When the CM is allowed to recover it is reported under a correct interface.
CMTS-Drivers
•
Symptom: Ironbus restarts have been observed on the Cisco uBR10012 due to ironbus link status 0x1180 errors. You observe the following messages in the PRE log: ------------------------------------------------------------------------------------------------------ 811468: Jun 17 03:10:17.233 UTC: slotindex is 8. 811469: Jun 17 03:10:17.233 UTC: IB Link status: 00001180 811470: Jun 17 03:10:17.233 UTC: %C10KEVENTMGR-1-IRONBUS_FAULT: Ironbus Event 5/0, Restarting Ironbus 811471: Jun 17 03:10:17.645 UTC: %C10KEVENTMGR-1-IRONBUS_SUCCESS: Ironbus Event 5/0, Restart Successful -----------------------------------------------------------------------------------------------------
The ironbus link status 0x1180 error will trigger an line card switchover on uBR10012s configured with N 1 redundancy. The ironbus restart is fast enough to keep modems online and has negligible affect to customers on uBR10012s without N 1 redundancy.
Conditions: This ironbus link status 0x1180 error has only been observed on slot 5/0 with the following hardware configuration. - Working MC520H Linecard in slot 5/0 (the problem has not been reported on the MC520U/S cards) - Active PRE2 in slot A (the problem has not been reported on a PRE1) The CMTS BU is still investigating the root cause of the ironbus link status 0x1180 error.
Workaround: There are two workarounds for ironbus link status 0x1180 error 1. Use Active PRE2 in slot B 2. Do not use a MC520H in slot 5/0 Customers should not RMA any equipment due to the ironbus link status 0x1180 error. The CMTS BU has not confirmed that this is a hardware issues and would need to identify the faulty hardware (e.g., PRE2, MC520H, chassis) if it is a hardware issue.
CMTS-Management
•
Symptom: When setting wide band interface to secondary through snmp, it doesn't work as expected.
Conditions: This symptom is observed in a wide band cmts which configure the wide band interfaces.
Workaround: Modify the wideband configuration through the CLI.
•
Symptom: Memory leak at buffer_peak_notify has been observed.
Workaround: There is no known workaround at this time.
CMTS-Packetcable
•
Symptom: When CM is primaryed on modular interfce, multiple vioce call tone were observed when bulk call is ongoing. Dropout time is about 4500~5000ms It is difficult to be reprodced on other testbed.
Workaround: There is no workaround.
CMTS-Platform
•
Symptom: When setting the Fragment-force 2000 the upstream performance drops to zero. There is no upstream thought-put.
Conditions: Problem is found only on the 5x20s card with ios ubr10k2-k8p6u2-mz.123-23.BC1.
Workaround: Put Fragment-force to Frag-force 1987 for best performance.
•
Symptom: PRE2 crashed multiple times at sch_rp_cm_reg_info_request_handler process.
Conditions: Above problem has been observed on ubr10K with PRE2 running Cisco IOS Release 12.3(23)BC4.
Workaround:There is no workaround.
•
Symptom: CMTS crashed at c10k_remote_crash_interrupt.
Conditions: The issue happens when continuously configuring 21BC customer configs.
Workaround: There is no workaround.
•
Symptoms: On a UBR10k the input rate counter of all GRE tunnels is incorrect. It's doubled compared to the real value. Only tunnels interfaces are affected.
Conditions: This occurs on a UBR10k/ESR-PRE2 running IOS versions 12.3(23)BC1 and 12.3(23)BC2 and 12.3(23)BC5.
Workaround: There is no workaround.
•
Symtoms: 1)Do not configure any fiber node. 2)Bring some narrow band modems online from SPA downstream. 3)Type "clear cable modem all reset" to reset modems. After this command, all the stat/BPI/PHS indexes in the modular-host tables are expected to be freed by design. But, this didn't happen. None of those indexes are freed. Consequence: This will cause stat/BPI/PHS indexes pool leaking, and leading to CMTS scalability issue, i.e., CMTS won't be able to support the stated amount of modems online from SPA downstream. Root cause: The current implementation for "clear cable modem all reset" is implemented for wideband modems, which checking the WB/NB association before releasing the indexes. This is not appropriate for a system which doesn't have fiber-node configuration, and is only to support NB modems.
Workaround: There is no workaround.
•
Symptoms: Modems go off line.
Conditions: pre-eq is enabled with ATDMA 64QAM 6.4Mz
Workaround: Use a different modulation profile.
•
Symptoms: From logs collected from the customer's network, at least three routers show the symptom of the problem. For the two SPAs installed in the Cisco Wideband SIP jacket card, SPA1/0/1 lost all its configurations, including its IP and MAC addresses, and all its RF channel configurations. SPA1/0/0 lost some of its configurations. IP and MAC addresses remain but some of the RF channel configurations are missing. Recovery measures were attempted, including PRE S/O, hw-module reset, unprovision/reprovision the SPAs, but none succeeded. Reloading the router seems to be the only way to recover.
Workaround: Reloading the router is the only way to recover.
•
Symptoms: Backup DTCC card may experience a unexpected reload irregularly with following messages. Oct 14 23:01:25.487: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:00:00:07 LKA:00:00:07 LRMV:never, NKA:00:00:00, cpu 1%/1% 2% Oct 14 23:01:25.491: %IPCOIR-3-TIMEOUT: KA Timeout: slot1/1, LRCV:never LKA:never LRMV:never, NKA:00:00:00, cpu 1%/1% 2% Oct 14 23:01:25.491: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is down. Notifying 2cable-dtcc driver. Oct 14 23:01:25.491: %UBR10K_ALARM-6-INFO: ASSERT CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1 Oct 14 23:01:26.555: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:26.555: %IPCOIR-5-CARD_LOADING: Loading card in slot 1/1 Oct 14 23:01:27.767: %C10K-5-LC_NOTICE: Slot[1/1] Line-card Image Downloaded...Booting... Oct 14 23:01:45.580: %IPCOIR-5-CARD_DETECTED: Card type 2cable-dtcc (0x5B0) in slot 1/1 Oct 14 23:01:45.656: %UBR10KTCC-4-CHG_CLK_REF: Clock reference source set to DTI for DTCC card 1/1 Oct 14 23:01:45.936: %UBR10K_ALARM-6-INFO: CLEAR CRITICAL slot 1/1/0 Active Card Removed OIR Alarm - subslot 1 Oct 14 23:01:45.936: %IPCOIR-2-CARD_UP_DOWN: Card in slot 1/1 is up. Notifying 2cable-dtcc driver. Oct 14 23:01:47.040: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active fast Oct 14 23:02:20.042: %UBR10KTCC-1-DTIPORTCLIENT: DTCC slot 1/1 port 1 DTI client status changed to Active normal
Conditions: This problem was seen on only Backup DTCC irregularly of uBR10K running 12.3(23)BC2.
Workaround: There is no workaround.
•
Symptoms: CLI shows 0-23 for ports, always 0-3 for subslot/bay.
Conditions: 12.3(23)BC, 12.2(33)SCA, 12.2(33)SCB, cobra_cable
Workaround: There is no workaround.
•
Symptoms: The CLC is crashing after the execution of the commands 'clear arp' and 'clear cable modem all delete'.
Workaround: No workaround.
•
Symptoms:When running 12.3(21a)BC7 and BC8, there are instances where the CPE device will disappear from the cable host table and sometimes from the ARP table. In all cases, the CPE loses IP connectivity. This shows up across different CM vendors, different CLC types, different plants, different 10ks. (This has also been seen in earlier IOS versions, including 12.3(17b)BC8. These are usually the symptoms seen: a) CPE loses connectivity. b) The CPE is not showing up on the CMTS using any "show" CLI like, "show cable modem <cm> <cpe>" or "show cable <interface> modem <sid>" c) Sometimes, the ARP entry for the CPE is seen in the ARP Table. d) When the CPE sends a DHCPDISCOVER packet, the CNR responds with DHCPOFFER. But for some reason, the CPE keeps sending DHCPDISCOVER. e) The work around is to clear the host by running the CLI "clear cable host". f) Although the CPE is not found in any of the "show" CLIs, "clear cable host" when ran for the first t
Guest
