July 25, 2012

Thank god switches don't have a screen

Strange… nearly exactly 152 days after a firmwareupgrade some HP ProCurve 6600ml-24G’s they rebooted again. Just 152 days before they rebooted after 152 days of successfull operation, that’s why I ran a firmwareupgrade on them.

Weird crashlog:

CRASHLogfileshow

...Crash Log File Header.......................... Product:   HP J9263A Name:      HP Switch E6600ml-24G Date:      Oct  8 2011 17:39:18 Build:     85 Version:   K.15.06.0008 Directory: /sw/code/build/btm(K_15_06) CPU:       PPC85XX crash rec index = 1 boot type    = Hard Boot willBootType = UNKNOWN BOOT .................................................. CrashRecordPointer (ffe4ea8)  for Crash Record Index 1 ----- Crash Record:   1 at 0xffe4ea8 ----- crash id = 0xeabba631 crash info = 0xbaabface subSystem ID = 0 timestamp:  05/15/12 04:09:48 Crash msg:  Software exception in kernel context at ghsException.c:1101 -> Internal system error

However, I didn’t find any caveats or fixes regarding this special
issue in the release notes of – yes that’s no typo – 11th of May. Just
four days before the crash. The HP support engineer told me that there
were “serveral versions coming out since that date”. Several versions…
Needless to say that they told me the same 152 days before the
spontaneous reboots after a nearly constant amount of time is fixed.
Obviously not… so we’re looking at our monitoring and keeping our
interns on hold for running through the datacenter.