Guest

Cisco MGX 8200 Series Edge Concentrators

Field Notice: CSCee20455 Causes Wrong Connections to Show A-Bit Alarms


May 12, 2004


Products Affected

Product

Comments

All FRSM Models

In MGX running 1.2.20 and 1.2.21 software

All CESM Models

In MGX running 1.2.20 and 1.2.21 software

All AUSM Models

In MGX running 1.2.20 and 1.2.21 software

Problem Description

Wrong connections go into A-Bit alarm at remote end.

This problem does not occur with the MGX-VISM or MGX-VISM-PR cards.

Background

When the prot resync mechanism gets triggered on an MGX running 1.2.20 or 1.2.21, the status of the wrong port is sent by the Service Module.

The net effect is that wrong interfaces programmed on the PXM1 get the port status from the service module. When a trigger event happens, the PXM1 sends A-Bit updates to the other end of the connections, and the wrong set of connections are put into alarm.

All data traffic on impacted ports stops. For a port to be impacted, there are two necessary conditions:

  1. On the same SM, there must be another port whose ID is exactly half of the port that is being impacted

  2. The state of this second port must be different than the state of the port being impacted. Example:

    For port #16 to be impacted, port 8 must exist (on the same SM) and the state of port#8 should be different than the state of port #16.

Once the above conditions are met, the following set of events must occur for the bug to surface:

  1. A switchcc, softswitch or port resync (port resync is a background task and happens every few hours). This will plant the problem but the data traffic will not be stopped at this point.

  2. A network event, such as a CPE line/port bounce, or any other event that will prompt the PXM to send A-bit updates toward the network has to happen. This will trigger the outage on all PVCs attached to the problem port.

Problem Symptoms

Connections show A-Bit alarms even though remote end is not in alarm. The result is that the attached Frame Relay CPE shuts down the connections that are incorrectly showing A-Bit alarms and traffic stops passing.

mgxnod1.1.7.PXM.a > dspcons -slot 1
     This End        Node Name       Other End       Status
                                                          
     1.1.0.40        mgxnod1        1.9.0.23        OK
     1.1.0.41        mgxnod1        3.19.0.23       OK
     1.1.0.42        mgxnod1        3.6.0.23        FAILED <====================
     1.2.0.34        mgxnod1        7.1.11.95       OK

Workaround/Solution

WorkAround:

None.

Solution:

  1. Upgrade software to 1.3.00 or later 1.3 release or 1.2.22 or later 1.2 release.

  2. Downgrade MGX software to a version earlier than 1.2.20.

Recovery:

Run the following command under shellConn of the FRSM card. Run this every time this problem is reported.

The command is send_port_traps .

send_port_traps function takes 4 parameters.

Parameter 1: Port Number This value is zero-based. ie For value of 50 in dspparifs and dspports , we need to pass 49 (ie, 50-1)

Parameter 2: Trap Number On FRSM-2CT3, FRSM-2T3E3 and FRSM-HS2/B card use 50381. This is fixed!

Parameter 3: Line Number The line number on which the port is added.

Parameter 4: PortDs0ConfigBitMap. This number of ds0 added on this port.

Parameter 1,3, and 4, i.e. Port No, Line num and PortDs0ConfigBitMap can be obtained from the dspport output.

For example, on port 50:

mgxnod1.1.5.VHS2CT3.a > dspport 50

     SlotNum 5 
     PortLineNum 50 <--- Line Number (Parameter 3)
     PortNum 50 <--- Port Number (Parameter 1)
     PortRowStatus Add 
     PortDs0Speed 64k 
     PortDs0ConfigBitMap(1stDS0) 0xffffff(1) <--- (Parameter 4) 
     PortEqueueServiceRatio n/a 
     PortFlagsBetweenFrames 0 
     PortSpeed 1536 kbps 
     PortLmiSigConfMethod Manual 
     SignallingProtocolType AnnexDUNI 
     AsynchronousMsgs UPD_UFS disabled 
     T391LineIntegrityTimer 10 sec 
     T392PollingVerificationTimer 15 sec 
     N391FullStatusPollingCounter 6 
     N392ErrorThreshold 3 
     N393MonitoredEventCount 4 
     EnhancedLmi Off 
     PortState Active 
     PortSignallingState No Signalling Failure 
     CLLMEnableStatus Disable

So the command turns out to be:

send_port_traps 49,50381,50,0xffffff

Thus, for the above example, you need to call the function from shellConn with following parameter:

mgxnod1.1.5.VHS2CT3.a > shellConn

     mgxnod1.1.5.VHS2CT3.a > send_port_traps 49,50381,50,0xffffff 
     send_port_traps 49,50381,50,0xffffff 
     value = 0 = 0x0

     Parameter1 49 (ie, 50-1 .. because its zero based.) 
     Parameter2 50381 (Port up trap.) 
     Parameter3 50 (line num) 
     Parameter4 0xffffff (as shown in dspport output.)

Note that replacing the cards will not resolve this problem.

DDTS

To follow the bug ID link below and see detailed bug information, you must be a registered user and you must be logged in.

DDTS

Description

CSCee20455 (registered customers only)

port resync on pop1 updates wrong port ids

For More Information

If you require further assistance, or if you have any further questions regarding this field notice, please contact the Cisco Systems Technical Assistance Center (TAC) by one of the following methods:

Receive Email Notification For New Field Notices

Product Alert Tool - Set up a profile to receive email updates about reliability, safety, network security, and end-of-sale issues for the Cisco products you specify.