crashing Itanium server

[Nyck_1]

One of my rx4640's crashed this morning for no real reason. After I got it back up and running I checked the /var/adm/syslog/syslog.log file and there was nothing much in there. Are there any other log files located anywhere else I could look at?

I've looked in /var/tombstones and /var/crash.

[Jeeshan]

look at the folder /var/adm/crash. if your system crashed, the crash dump is generated here.

moreover you can check the MP logs for more info and also /var/opt/resmon/event.log

[Nyck_1]

There are a couple of crash dump directories but this server is currently not on support so I'm unable to log a call with HP.

The event.log is not picking anything up for the time of the crash, oh well cheers for the response.

[Ivan Krastev · Accepted solution]

Can you post the INDEX file from the latest dump?

regards,
ivan

[Patrick Wallek]

The current /var/adm/syslog/syslog.log will NOT have anything regarding the crash. The log file gets recreated when the system boots. The old file gets copied to /var/adm/syslog/OLDsyslog.log, so have a look in that file.

Also have a look in /etc/shutdownlog and see if anything interesting shows up there.

[Nyck_1]

Sure can, details below:-

more tmpindex
comment savecrash crash dump INDEX file
version 2
hostname lonita2
modelname ia64 hp server rx4640
panic Bad News!

dumptime 1215512386 Tue Jul 8 11:19:46 BST 2008
savetime 1215512856 Tue Jul 8 11:27:36 BST 2008
release @(#) $Revision: vmunix: B11.23_LR FLAVOR=perf Fri Aug 29 22:35:38 PDT 2003 $
memsize 34328862720
chunksize 268435456
defcompchunk 0x0000000000000000

virtual to physical information for IA64
registers 0x8d2cab0
vhpt 0xe000000180000000 8388608
itr 0x8d4fd20 63
dtr 0x905f000 65

module /stand/crashconfig/vmunix vmunix 64260968 1026076136
module /stand/crashconfig/mod/rng rng 92744 4252986015
module /stand/crashconfig/mod/dmpaa dmpaa 20384 2270847948
module /stand/crashconfig/mod/dmpaaa dmpaaa 21160 1064655384
module /stand/crashconfig/mod/dmpap dmpap 20808 1051718878
module /stand/crashconfig/mod/dmpapg dmpapg 25080 271114488
module /stand/crashconfig/mod/dmpapf dmpapf 27216 549770395
module /stand/crashconfig/mod/dmpjbod dmpjbod 20440 2589002619
module /stand/crashconfig/mod/dmphpalua dmphpalua 28064 2861264650
module /stand/crashconfig/mod/dmphdsalua dmphdsalua 25528 1988916574
module /stand/crashconfig/mod/pfil pfil 134592 3574228516
module /stand/crashconfig/mod/ipf ipf 583824 4051795845
module /stand/crashconfig/mod/gvid_info gvid_info 30664 2844918749
module /stand/crashconfig/mod/mvfs mvfs 2275024 2277021947
image image.1.1 0x0000000000000000 0x0000000000031000 0x0000000000000000 0x000000000000009f 2968254122
image image.2.1 0x0000000000000000 0x000000000cd94000 0x0000000000000100 0x000000000003df23 3301013230
image image.3.1 0x0000000000000000 0x0000000000061000 0x000000000003e23c 0x000000000003e527 259320949
image image.4.1 0x0000000000000000 0x0000000000061000 0x000000000003fc00 0x000000000003fe1f 1696440113
image image.5.1 0x0000000000000000 0x000000000fffa000 0x0000000000100000 0x000000000010fff7 3970609185
image image.5.2 0x0000000000000000 0x000000000fffa000 0x000000000010fff8 0x000000000011ffef 3078342080

[Nyck_1]

yeah I looked in the OLDsyslog.log file first as that is where the issue would of been listed.

[Ivan Krastev]

Check for this patch PHKL_34666 - http://www11.itrc.hp.com/service/patch/patchDetail.do?patchid=PHKL_34666&sel={hpux:11.23,}&BC=main|search|


Also try to analyze crash dumps for more information.

regards,
ivan

[Nyck_1]

I've just checked and that patch is already installed.

When it comes to checking the crash dumps, what commands would you use?

[Ivan Krastev]

See this previous thread - http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1089090

There is also a tool called crashinfo, attached in one of the posts.

regards,
ivan

[Nyck_1]

nice one, that crashinfo tool rings a bell from somewhere in the distance past:-)

[Don Morris_1]

If you do get crashinfo, post the resulting output please. If not, at least go into the crash dump directory and do:

kwdb -q4

trace -u event 0

examine &msgbuf+0x8 using s

Capture/post that output [crashinfo's output will include this stuff, so don't include both -- the critical things here are the message buffer right before the crash and the stack trace of the crash itself... that's the starting point for analysis in most cases].

[Nyck_1]

I have found the crashinfo program so I will post results when I have had the chance to run it

[Robin T. Slotten]

Here is a good document about the crash dumps.

Rob...