8212zl : Software exception in ISR at btmDmaApi.c:378

[jc_paris]

Hi

Yesterday, I had a problem with a switch 8212zl. MM1 reset at 14h25 and MM2 became active. At 18h25, switch reload and MM1 became active. For this error, I have this error : 

VIT-SERV-R253# sh boot-history

Mgmt Module 1 -- Saved Crash Information (most recent first):
=============================================================
Mgmt Module 1 in Active Mode went down:  08/18/11 14:25:33
Software exception in ISR at btmDmaApi.c:378
-> ASSERT:  No resources available!

Mgmt Module 1 in Active Mode went down:  05/03/11 15:27:52
Operator warm reload.

Mgmt Module 2 -- Saved Crash Information (most recent first):
=============================================================
Mgmt Module 2 in Active Mode went down:  08/18/11 18:44:59
Software exception in ISR at btmDmaApi.c:378
-> ASSERT:  No resources available!

No Core-dump Files Present.
VIT-SERV-R253#

What does this error ?

[cenk sasmaztin]

hi

two MM to have same software version

[cenk sasmaztin]

check this process in the link

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c02905316&jumpid=em_alerts_us-us_Jul11_xbu_all_all_1390686_90650_networking_routine_013_775

[Helper]

Hi,

What is your actual firmware version ?

Do you have a 'show log -r -a' output to take a llok to the events during the issue ?

Bye,

[Helper]

As per the RN documentation to see all the logs, from active and standby MGMT module you should add '-s' to the show log -r -a.

For your information :
Version K.15.03.0006

■ Crash (PR_0000066570) —After a large number of startup configuration changes, the switch might reboot
unexpectedly with a message similar to the following.
Unable to allocate message buffer
Software exception in ISR at btmDmaApi.c:370


Version K.15.04.0002


■ Module Crash (PR_0000064847) —A switch module might reboot unexpectedly with a message similar to the
following.
Software exception in ISR at buffers.c:3222
-> ASSERT0: failed

[jc_paris]

Hi

Here are the firmwares used and the log (see attachment): 

VIT-SERV-R253# sh version
Management Module 1: Active
Image stamp:    /sw/code/build/btm(K_15_04)
                Mar  8 2011 16:16:33
                K.15.04.0003
                56
Boot Image:     Primary

Management Module 2: Standby
Image stamp:    /sw/code/build/btm(K_15_04)
                Mar  8 2011 16:16:33
                K.15.04.0003
                56
Boot Image:     Primary
VIT-SERV-R253#

On release note, I also saw the crash but it happens on older firmwares.

[Helper]

Hi,

It miss some logs before the first failover to have a better understanding of the situation :

W 08/18/11 18:45:19 00268 system: xM1: Mgmt Module 1 - Failover occured
E 08/18/11 18:45:12 02648 srcip: AM2: SYSLOG - failure to send out pkt

The syslog failure message do not give us much explanations about the behaviour.

We can see "many" run to start backup after the initial failover and until the normal situation.

I 08/18/11 18:45:20 02617 mgr: AM1: Startup configuration changed.  New seq.
            number 123

I 08/18/11 18:45:49 02612 mgr: AM1: chassis subsystem saved the whole running config to startup config. I 08/18/11 18:45:49 02612 mgr: AM1: chassis subsystem saved the whole running config to startup config. I 08/18/11 18:45:49 02612 mgr: AM1: chassis subsystem saved the whole running config to startup config.

I 08/18/11 18:45:57 02612 mgr: AM1: chassis subsystem saved the whole running
            config to startup config.

I 08/18/11 18:45:58 02617 mgr: AM1: Startup configuration changed by SNMP. New seq. number 124

I 08/18/11 18:46:19 02617 mgr: AM1: Startup configuration changed.  New seq.
            number 125
I 08/18/11 18:46:19 02611 mgr: AM1: chassis subsystem saved some internal
            change(s) to startup config.

I 08/18/11 18:53:39 02617 mgr: AM1: Startup configuration changed by SNMP. New seq. number 126

There are some events that need to be analyze.

W 08/18/11 18:47:21 00374 chassis: AM1: Fabric Event F0: SLOT A NCI #7 -
            NCI-A:INT1 0x08000800
W 08/18/11 18:47:19 00374 chassis: AM1: Slot D: Lost Communications detected -
            Source Message System(40)
I 08/18/11 18:47:19 00076 ports: AM1: port A7 in Trk2 is now on-line
I 08/18/11 18:47:19 00435 ports: AM1: port A7 is Blocked by STP
I 08/18/11 18:47:18 00435 ports: AM1: port A7 is Blocked by LACP
I 08/18/11 18:47:18 00832 udld: AM1: Link state on port 'A7' is changed to 'up'
            state.
W 08/18/11 18:47:17 00374 chassis: AM1: Slot B: Lost Communications detected -
            Source Message System(40)
I 08/18/11 18:47:13 00435 ports: AM1: port A7 is Blocked by UDLD
I 08/18/11 18:47:13 00832 udld: AM1: Link state on port 'A7' is changed to
            'failure' state.
W 08/18/11 18:47:05 00374 chassis: AM1: Slot D: Msg loss detected - no ack for
            seq # 3709
W 08/18/11 18:47:05 00374 chassis: AM1: Slot B: Msg loss detected - no ack for
            seq # 3684

One suggestion regarding this events

W 08/19/11 10:21:55 00236 snmp: AM1: SNMP Security access violation from
            10.1.210.2

You should modify the SNMP credential used to contact your device or remove snmp polling from this end-node.

At that time i can't further help you, it should be intersting to open a case. If you have events logging to an external server, maybe snmp traps, syslogs or radius/tacacs accounting could help us to clarify this situation.

Bye.