Procurve Manager 3.10 crashes Procurve 2610 switches

[Ole Holm Nielsen]

We have installed Procurve Manager (latest release with updates: PCM C.03.10.201) on a WinXP PC a few days ago. After installing PCM we suddenly have two independent Procurve 2610 switches crashing every night at exactly 02:01 a.m. This has not happened before. The 2610 switches run the latest firmware R_11_48 and show this in the event log at the time of crash:
sys: 'TLB Miss: Virtual Addr=0x53595358 IP=0x806e9234 Task=
PCM shows:
Event Description
ColdStart - Device has crashed or power plug is removed or SNMP entity is re-initializing itself.

Questions:
1. How can we prevent PCM from doing a task at 02:01 that causes crashes (we did not configure any tasks).
2. How can we report this error to HP ?

Thanks,
Ole

[Michael_Breuer]

Hello Ole,

I suggest the following:

1) exclude this device from the discovery that the PCM doesn't touch this switch at all (you find this option in the agent configuration)
2) Collect sh tech information and open a ticket at HP support

Cheers,

Michael

[Ole Holm Nielsen]

Thanks to Michael_Breuer for excellent advice. I had however made a different attempt before I read his answer: I downgraded 1 of the 2 Procurve 2610 switches from R_11_48 to R_11_30. The downgraded switch DOES NOT CRASH. The switch running R_11_48 crashes at 02:01 a.m., just at the time when PCM reports under the device Configuration view: "Last scan: 03/12/2010 02:00:10".
So the conclusion seems clear: Procurve 2610 firmware R_11_48 crashes due to PCM scanning it for configuration information ! I have now downgraded both of these switches to the older firmware R_11_30.
Question: This problem appears to be due to a bug in the Procurve 2610 firmware, but what is the correct channel for reporting this to HP Procurve ? I have no experience in reporting such software bugs.

[Ole Holm Nielsen]

Note added: It seems that HP has withdrawn Procurve 2610 firmware release R_11_48 from the download page http://www.procurve.com/customercare/support/software/switches.htm.
I haven't been able to find any information on the website as to the reason for the withdrawal :-(

[Ralf Krause · Accepted solution]

This in fact is a bug in R.11.48 through R.11.50 releases.

They already had been removed from the Web and messages should have been sent to those having downloaded the releases form the official web page.

To report this issue, call ProCurve support:
https://my.procurve.com/help/help_topics.aspx?CultureCode=en-US

[Shadow13]

Downgrade to R.11.30 for now and the scan should work normally with no issues.

[Marco Albertini]

Same problem for me, with Procurve Manager Plus 2.3 and 2610 switch with R_11_48 firmware.

I got many different error messages:

ProCurve J9088A Switch 2610-48
Software revision R.11.48

switch-2# sh boot-hist

Master -- Saved Crash Information (most recent first):

======================================================

SubSystem 0 went down: 04/02/10 15:37:09
Unalligned Access: Virtual Addr=0x00000025 IP=0x806e9234 Task='mIpCtrl'
Task ID=0x85d27ec0 fp:0x00000000 sp:0x85d27d78 ra:0x806d06ec sr:0x1000fc01

SubSystem 0 went down: 04/02/10 15:25:24
TLB Miss: Virtual Addr=0x20000004 IP=0x806e9234 Task='tSvcWorkQ'
Task ID=0x85972b00 fp:0x00000000 sp:0x85972a10 ra:0x806d06ec sr:0x1000fc01

SubSystem 0 went down: 04/02/10 13:08:24
TLB Miss: Virtual Addr=0x53595358 IP=0x806e9234 Task='mftTask'
Task ID=0x85962fe0 fp:0x85962600 sp:0x85962538 ra:0x806d06ec sr:0x1000fc01

SubSystem 0 went down: 04/02/10 12:30:29
Unalligned Access: Virtual Addr=0x00000025 IP=0x806e9234 Task='tSvcWorkQ'
Task ID=0x85972b00 fp:0x00000000 sp:0x85972a10 ra:0x806d06ec sr:0x1000fc01

SubSystem 0 went down: 03/28/10 03:01:11
TLB Miss: Virtual Addr=0x20000004 IP=0x806e9234 Task='tSvcWorkQ'
Task ID=0x85972b00 fp:0x00000000 sp:0x85972a10 ra:0x806d06ec sr:0x1000fc01

[Massimo Poletti_1]

I have 45 2610-24/48 switches in my company running fw 11.48.

Every night a scheduled task scan the configurations to look for changes.

No problem at all until now, but I will upgrade this (seems to be) buggy firmware as soon as the brand new will be downloadable!

I agree that a brief note from HP in download page explaining why 11.48 has been removed could be very useful for all Procurve customers like me, asking why the last version is now 11.30!

[Marco Albertini]

Hi Massimo,
nice to hear that someone doesn't have our problems.

Some questions:

- Which scheduled task do you use ? Is it "default scan device configuration" or "ARP/Device attributes/Neighbor discovery" ?

- Do you use SNMPv3 ?

- Is tftp client/server enabled on switches ?

Thank you in advance
Regards,
Marco

[Massimo Poletti_1]

Hi Marco.

Damn...

I looked in the wrong place.
The devices does crash!

Since doesn't matter for me a restart in the middle of the night I will not downgrade the firmwares.

Anyway, my answers:
- Default scan device configuration
- No
- Yes

Ciao
Massimo