Guest

Cisco MGX 8230 Software

After Upgrading From MGX Release 2.1 to MGX 3.0.20 and Earlier Releases Results in Database Anomalies


June 18, 2003


Products Affected

MGX8800

Problem Description

After an upgrade from MGX release 2.1 to MGX 3.0.20 and earlier releases, some of the ports partition database information on the ATM Switching Service Module (AXSM) cards is corrupted. This is manifested only when AXSMs are in redundancy mode and if multilple partitions, for example Private Network-to-Network Interface (PNNI) and Multiprotocol Label Switching (MPLS) are configured.

Background

The 3.0.x code of AXSM introduced the dynamic partition feature. This essentially enables a partitions BW/LCN/VPI-VCI to be shrunk or extended depending on the needs of the customer. MGX Release 2.1.x does not support the dynamic partition feature. It does support multiple partitions. Partitions segregate the card resources such as bandwidth, logical channel numbers (LCNs), virtual path identifiers (VPIs) and Virtual Channel Identifiers (VCIs) for use by either a PNNI controller or an MPLS controller. The user must configure partitions when planning to mix PNNI and MPLS services in a single network. Appropriate resource partitioning is important for scaling each service type. Corruption of the partition database will result in anomolous behaviors such as failure to commit connections and incorrect load displays.

Problem Symptoms

Tag lines will be seen in PNNI partition.

MPLS partition will probably be empty. Inspect the exceptions portion of dspload ifNum partId for each partition.

Sample of corruption from dspload. This is a partial display showing only the information of interest

+--------------------------------------------+
| E X C E P T -- V A L U E S |
+--------------------------------------------+
| SERV-CATEG | VAR-TYPE | INGRESS | EGRESS |
| VBR-nRT | Avl Chnl | 0000008 | 0000008 |
| UBR | Avl Chnl | 0000008 | 0000008 |
| ABR | Avl Chnl | 0000008 | 0000008 |
| TAG_COS_0 | Avl Chnl | 0000008 | 0000008 |
| TAG_COS_1 | Avl Chnl | 0000008 | 0000008 |
| TAG_COS_2 | Avl Chnl | 0000008 | 0000008 |
| TAG_COS_3 | Avl Chnl | 0000008 | 0000008 |
+--------------------------------------------+
| VBR-nRT | Avl Bw | 0002351 | 0002351 |
| UBR | Avl Bw | 0002351 | 0002351 |
| ABR | Avl Bw | 0002351 | 0002351 |
| TAG_COS_0 | Avl Bw | 0070501 | 0070501 |
| TAG_COS_1 | Avl Bw | 0000001 | 0000001 |
| TAG_COS_2 | Avl Bw | 0282001 | 0282001 |
| TAG_COS_3 | Avl Bw | 1057504 | 1057504 |
+--------------------------------------------

Note: In a partition display that is not corrupted it will not have ATM and tag information in the same partition. That is, Variable Bit Rate (VBR), Unspecified Bit Rate (UBR) or Available Bit Rate (ABR) connection types should not appear in the same partition as MPLS connection types. MPLS connection types are shown a TAG_COS_n. Following is an example of an uncorrupted dspload screen capture.

+--------------------------------------------+
| E X C E P T -- V A L U E S |
+--------------------------------------------+
| SERV-CATEG | VAR-TYPE | INGRESS | EGRESS |
+--------------------------------------------+
| TAG_COS_0 | Avl Bw | 0070501 | 0070501 |
| TAG_COS_1 | Avl Bw | 0000001 | 0000001 |
| TAG_COS_2 | Avl Bw | 0282001 | 0282001 |
| TAG_COS_3 | Avl Bw | 1057504 | 1057504 |
+--------------------------------------------+  

Workaround/Solution

It is recommended to not upgrade to MGX 3.0.20 if the network environment has multiple partitions on the same ports. Monitor bug ID CSCea55999 for a resolution. It is expected but not guaranteed that this anomaly will be resolved in MGX Release 3.0.23 and later releases. In the case where the MGX8800 has been upgraded from 2.1 to 3.0, follow the procedure below to clean up the corrupt data.

Workaround:

  1. Reset both AXSMs in the redundant pair so that both end up booting in the same timeframe. This method is not the most graceful way.

  2. Partition recovery procedure shown below. This is graceful if bandwidth is available.

Before starting, capture the following information:

For the AXSM pair:

  • dspports

  • dspparts

For each port:

  • dspport

For each part:

  • dsppart

  • dspload

For the LSC:

  • show ip interface brief

For each XtagATM int:

  • show interface

Begin:

Login to the MGX

telnet

Login: cisco

Password:

*** Recursion Entry Point ***

Go to the LSC and shut the XTagATM interface for the first port to force MPLS traffic to route around this interface. It may also be desirable to work on the other end of this trunk at the same time.

Commands:

  • cc 3

  • configuration terminal

  • interface XTag

  • shutdown

  • control z [^z]

Go to the AXSM card pair. cc and choose the active card.

Down the first port.

  • dnport

Delete the two partitions on the port.

  • delpart

Delete the port.

  • delport

Add the port back.

  • addport

HINT: Cut and paste from dspports capture, and add iftype 2.

Add back the two partitions.

  • addpart

HINT: Cut and past everything after addpart from dspparts capture.

Return to the LSC and up the XTag interface

Commands:

  • cc 3

  • configuration terminal

  • interface XTag

  • no shutdown

  • control z [^z]

For all remaining interfaces return to the *** Recursion Entry Point*** until all interfaces have been done.

Verification and Post Check

For the AXSM pair:

  • dspports

  • dspparts

For each port:

  • dspport

For each part:

  • dsppart

  • dspload

Verify that the proper traffic types appear.

For the LSC:

  • show ip interface brief

For each Xtag interface:

  • show interface

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

CSCea55999 (registered customers only)

dspload shows service category corruption for partitions

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.