The Standby CP on Brocade DCX Switches Might Unexpectedly Reboot Because of the Time Service Daemon ( tsd ).

Asset ID:	1-72-2133515.1
Update Date:	2016-05-05
Keywords:

Solution Type Problem Resolution Sure

Solution 2133515.1 : The Standby CP on Brocade DCX Switches Might Unexpectedly Reboot Because of the Time Service Daemon ( tsd ).

Applies to:

Brocade DCX Backbone - Version All Versions to All Versions [Release All Releases]
Brocade 48000 Director - Version All Versions to All Versions [Release All Releases]
Brocade DCX-4S Backbone - Version All Versions to All Versions [Release All Releases]
Information in this document applies to any platform.

Symptoms

The switches affected by this issue are Brocade DCX, DCX-4S and 48000 running FOS version 6.4.0 or earlier.

Using the new timezone format ( country / city format ) may lead to the time service daemon ( tsd ) running out of memory. The key symptom of this issue is
the following message observed on the RASlog file:

2016/04/20-05:52:29, [KSWD-1002], 39767, SLOT 7 | FFDC | CHASSIS, WARNING, COLDCX1, Detected termination of process tsd:2005
Service instances out of sync

Followed by a subsequent reboot of the standby control processor ( on the example below it’s inserted on chassis slot 7 ).
A FFDC event is generated.

2016/04/20-05:52:29, [FSSM-1003], 39768, SLOT 6 | CHASSIS, WARNING, COLDCX1, HA State out of sync.
2016/04/20-05:52:29, [RAS-1002], 39769, SLOT 7 | CHASSIS, WARNING, COLDCX1, First failure data capture (FFDC) maximum storage size (4 MB) was reached.
2016/04/20-05:52:29, [RAS-1001], 39770, SLOT 7 | CHASSIS, INFO, COLDCX1, First failure data capture (FFDC) event occurred.
2016/04/20-05:52:29, [HAM-1014], 39771, SLOT 7 | CHASSIS, CRITICAL, COLDCX1, Non restartable component (ts (pid=2005)) died.

2016/04/20-05:52:29, [FW-1424], 39772, SLOT 6 | FID 128, WARNING, COLBrocade-1, Switch status changed from HEALTHY to MARGINAL.
2016/04/20-05:52:29, [FW-1433], 39773, SLOT 6 | FID 128, WARNING, COLBrocade-1, Switch status change contributing factor CP: CP non-redundant (Slot7/CP1) faulty.
2016/04/20-05:53:34, [EM-1033], 39774, SLOT 6 | CHASSIS, ERROR, COLDCX1, CP in Slot 7 set to faulty because CP ERROR asserted.

Cause

This is caused by Brocade defect Defect000319033

Solution

For a permanent solution, the switch needs to be upgraded to FOS v6.4.1a or higher.

For a workaround, users might switch to the old timezone format, using the tstimezone command.

Syntax:

tstimezone --old <hours> [,] [minutes]

For example, the command

tstimezone --old 0,0

would set timezone to GMT.

Note: After the switch has been upgraded to the recommended versions, you need switch back the timezone to the new format ( country / city ),
since the old one doesn’t set Daylight Saving Time adjustments.

To check which Oracle patches contain the necessary firmware versions to get this fix, please refer to the following document:

Oracle Fibre Channel (FC) SAN Switch Firmware Patch Matrix for Firmware Upgrade - Brocade - Cisco - McData - Qlogic (Doc ID 1507746.1)

References

<NOTE:1507746.1> - Oracle Fibre Channel (FC) SAN Switch Firmware Patch Matrix for Firmware Upgrade - Brocade - Cisco - McData - Qlogic

Attachments

This solution has no attachment