System Hangs

[Halan Leno Borges Dias]

Hello...

I've a N4000 with service guard Cluster, yesterday one of this servers crashed. I'd like to find out what happend. What log file I must check to verify the causes of server crash ?

Regards
Halan Leno
Brazil

[T G Manikandan]

check the /var/adm/syslog/syslog.log

for errors.
If you had rebooted the system then check the OLDsyslog.log file.

ALso check the directory /var/adm/crash for crash files

[Chris Wilshaw]

You could check

/var/adm/syslog/OLDsyslog.log

/etc/shutdownlog

/var/tombstones/ts99 (This will list chassis codes if there is a hardware problem, which an HP engineer could use to find out what was at fault).

[Armin Feller]

Hi,

please check following:

on chrashed Node:
1.)
# cd /var/adm/crash/crash.x
# more INDEX
---> check reason for panic

# . /usr/contrib/Q4/bin/set_env
# q4pxdb
# q4 vmunix
---> and debug the dump, if you have a contract with HP call them for help.

2.)
# more /var/adm/syslog/OLDsyslog.log
---> check here for a reason, such like "cmcld died", ...


on the other Nodes:
# more /var/adm/syslog/syslog.log
---> check here also for a reason


Regards,
Armin

[Ravi_8]

Hi

/var/adm/syslog/OLDsyslog.log and
/var/adm/crash will give you the cause in detail

[Armin Feller]

please check if patch:

10.20 PHCO_21061
11.00 PHCO_21062
11.11 PHCO_25723

is installed. If you can do following:

# . /usr/contrib/Q4/bin/set_env
# q4pxdb
# q4 vmunix

q4> run WH>wh.out
q4> quit

# more wh.out
---> here you will find al lot of informations, search for reason of panic


IF MC/SG toc'ed the system then you have to search in OLDsyslog for more details.

[V. Nyga]

Hi,

maybe this would be interesting for you (Crash of N4000er!):
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xe05d0fe6d0f7d61190050090279cd0f9,00.html
and a discussion of that:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xeb9c3a1c04ffd61190050090279cd0f9,00.html

Regards
Volkmar

[V. Nyga]

P.S.

and don't hesitate to give points if you guess the guys here in the forums maybe helped you...
country: BRAZIL
personal quote:Halan Leno
ITRC member since: July 17, 2002
last contribution date: November 28, 2002

This member has assigned points to 0 of 20 responses to his/her questions.

We are pleased if we can help you!
Volkmar

[Bill McNAMARA_1]

my guess is that the reason for the shutdown is a TOC, ie SG initiated.
Your syslog of all nodes should give you more.

See SG docs:
http://docs.hp.com/hpux/ha/index.html#ServiceGuard