Solaris

I have an oracle cluster management and noticed that when Oracle evict a node (Sun Solaris 10 server); it does cause a reboot that looks like a power outage (reference to 3320 Disk Array Configuration for Oracle 10) that is not logged in the messages log. After further investigation, it was found out that Oracle is issuing "/sbin/uadmin 1 1". The man page for uadmin is not very useful so I checked the internet and found out that "uadmin 1 1" means do not sync filesystems and reboots to multi-user mode. I have tested "uadmin 1 1" and "uadmin 2 1" (which means sync the filesystems and reboots to multi-user mode) and noticed that when in both cases the messages log does not show who issued the command (like what you see when you issue "reboot")! In the case of "uadmin 2 1" the last thing the messages log shows is "syncing file systems..." and then nothing until the server comes back up.

I have contacted Sun and the support claimed that when uadmin is issued it should be logged in the messages log but after some time of discussion he agreed that the messages log do not log it but does not know why.

Anybody know more details on this command, what is the different between it and "reboot" and why it is behaving like this if this is the proper behavior?

I have already considerable amount of time searching so any response is might be helpful and appreciated.

This is not extrange, the reboot command have options to avoid the generation of logs of the reboot event. This options are
-n or -d , from the man

------------
The reboot utility normally logs the reboot to the system log daemon, syslogd(1M), and places a shutdown record in the login accounting file /var/adm/wtmpx. These actions are inhibited if the -n or -q options are present.
--------------


One difference between uadmin and reboot is that with the uadmin command you can poweroff the system, for example "uadmin 2 6", with reboot you can't make this.

Additionaly , with reboot you can send options for the boot process, for example "reboot -- -r" to reconfigure, or "reboot -- second_disk" to indicate to the system that second_disk is the boot device, with uadmin I'm not sure that it could be posible.

I hope that it be useful.

Thanks for the information cerber0! So if uadmin 1 1 is issued, is it considered graceful or ungraceful shutdown? In other words, does this action stops processes gracefully and properly or not? If not, then I guess it might cause disk/OS corruption!!??