Solaris

hi,
how do i find boot reason in solaris?
i know that signal 15 reboot is a manual reboot...
#cat /var/adm/messages | grep -i signal

what does this mean from #last -10

"reboot system boot Sun Oct 19 09:24"

is it a system boot (auto reboot by the server) or human reboot?

how to differentiate human/server boots?

please advise...

related command i use to check the boot info

#who -b
#last -10
#last|more
#last reboot|more

what type of hardware are you using??? a system with a lom or a system controller can tell you these things. the larger system need to be "setswitchkey off" in order for a complete reboot of that particular system. something with a lom say a netra t1 will need to be 'poweroff' which you can view with the show eventlog command.

hi pupp,
these are netra and Enterprise 250 (UltraSPARC-II 296MHz) servers.
but am not sure of the lom / systems controller as i am still a junior sys ad.
what i would like to is how to differentiate human/server boots?

#dmesg | grep -i panic
#dmesg | grep -i boot
#dmesg | grep -i signal

are these sufficient...please advise..thanks.

no worries.

what are you intentions for this??? the reason i ask is if you are simply trying to monitor these devices... use opennms or openview. these can help spot when a server goes down as a result of a panic/fault or a human initiated reboot/shutdown. but just because a panic happens does not mean the system will reboot. just recently we had a system reboot itself (it was seen on OV). the system did not panic. in fact it was a kern.info. and then the system came down on signal 15 (a simple exit and no core dump). on another box we actually rebooted it. again, signal 15. so i think it might be hard to differentiate between a human reboot and a fault induced reboot. also, if the signal permits, a core dump will be generated giving you a good idea that the system just crashed.

also, i believe a proper shutdown should terminate syslogd correctly. this will say its coming down on signal 15. if you look in your logs and find the output of `uname -a` in the log files and a gap in your log files... most likely it was a crash/reboot or something a long those lines.

the reason is,at my work place we have 1000 over servers..and at times we receive incidents relating to "boot detected" alerts and we have to find the reasons.
so am still new in this area, at times itz hard diff between human and machine reboot.
anyways,thanks alot for you valuable info.greatly appreciate them.you have a superb day! thanks again

Agree with pupp, sometimes if there were to be a sec of power-surge/outage, the system would also reboot and the only thing you can see is a signal 15 in your messages file. In older systems, I have encountered such incident and you wont find any errors in your system and would find it rebooting intermitently. WE COULDN'T FIND ANY PROBLEMS EVEN IN THE PRTDIAG OUTPUT. But after replacing the PS unit, all went fine. Sun servers... wierd souls..