Hi,
I am running Solaris 2.6, latest patches, across about 80 systems and
experiencing the occasional system panic... usually explainable, but
sometimes not.
The problem is that the crash dumps are not being written to
/var/adm/messages... but they are being written to the console.
eg: The following appeared on the console during a recent system failure,
but not in the messages file:
panic[cpu0]/thread=0x30023e80: CPU0 Multiple Privileged UE Error:
AFSR 0x00000001 80300000 AFAR 0x00000000 09a00030 SIMM U09
01 U1001
syncing file systems... 4 done
6091 static and sysmap kernel pages
30 dynamic kernel data pages
254 kernel-pageable pages
0 segkmap kernel pages
0 segvn kernel pages
0 current user process pages
6375 total pages (6375 chunks)
excerpt from my syslog.conf:
*.err;kern.notice;auth.notice /dev/console
*.err;kern.debug;daemon.notice;mail.crit /var/adm/messages
Which syslog faciltiy/level does a cpu panic fall into ?
If I change the syslog.conf ( and HUP the sysogd) how can I simulate a panic
... and test the resultant logging ?
Thanks for any advice.
- kelvin.