Access via stack master crashes 2510-24

[T. Joebstl_1]

Hello,

I just upgraded the firmware to Q.10.05 on the only 2510-24 I have stacked together with some 2524's (running F.05.59).
The moment I try to access the 2510-24 from the stack master (one of the 2524's) either via telnet or the webinterface the 2510 crashes and restarts.

Reported error message is:
SubSystem 0 went down: 03/28/07 17:25:42
Unalligned Access: Virtual Addr=0xa3fc3909 IP=0x802bf8a8 Task='eDrvPoll'
Task ID=0x83511340 fp:0x00000002 sp:0x83511240 ra:0x80094b78 sr:0x1000fc01

Only fix is to either revert to Q.10.01 or disable stacking.

Anyone else seeing such issues?

Regards,
Thomas

P.S. This thread has been moevd from Switches, Hubs, Modems (Legacy ITRC forum) to ProCurve / ProVision-Based. - Hp Forum MOderator

[T. Joebstl_1]

Seems it gets even better...
Just got a 2510-24 in so I thought I'll play around with that and my spare 2424.

If the 2424 is the stack master, the 2510-24 crashes the moment I try to access it from the 2424.
If the 2510-24 is the stack master, it crashes the moment I try to access the 2424 via the login selection.

Same error message, same resolution.
Unfortunately I dont have a second 2510-24 or a 26xx here (the other one is at a remote site) else I'd try that as well.
I guess stacking - at least with 'old' devices - is severely broken in Q.10.05.

I'd still be interested if someone else sees such issues though.

Regards,
Thomas

[T. Joebstl_1]

Oh boy, I really start to wonder if HP is doing some interop testing between their own devices. The 2810-48G I just got in seems to have the same issues with N.10.07, going back to N.10.02 fixes this.

SubSystem 0 went down: 01/01/90 00:00:54
Unalligned Access: Virtual Addr=0xa3e4ce95 IP=0x802e7088 Task='eDrvPoll'
Task ID=0x81abbae0 fp:0x00000002 sp:0x81abb9e0 ra:0x800a7068 sr:0x1000fc01

[timlang]

I have the exact same problem with my 2810-24G switch. I had the 2810-24G setup as stack commander, with a 2626 as stack member. When I upgraded the firmware in my 2810 from N.10.02 to N.10.07, any time I try to do any sort of stack operations through the web browser, the 2810 has a hard crash with the same Unalligned Access error. I don't really need my two switches stacked, so I just disabled the stacking to avoid further problems. The ironic thing is that in the release notes for N.10.07, it says that N.10.06 fixed:

* Crash (PR_1000386489) â The switch may crash with a message similar to: Unalligned Access: Virtual Addr=0xa3e2dad7 IP=0x801f8234
Task='eDrvPoll' Task ID=0x81b02650 fp:0x81ecce90 sp:0x81b02538 ra:0x801f8208

so either they re-broke it in 10.07 or it is a different problem with a similar crash error.

Either way, I think the solution currently is either disable stacking or roll back to N.10.02