| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: 10g ASM Startup After Reboot
Hi Andy,
Sorry for not helping you directly.
But ugly thing about ASM and CSS is that 10g just
providing more complexity then before.
The startup procedure now is harder and very hard to
control.
Oracle said some ASM problems and difficulties they are going to resolve in Oracle 10g Release 2 that is maybe coming next month.
I hope they will sort out that CSS/ASM/DB crap :)
Who said no need for DBA's when 10g is out.
It is just more complicate now with more features then
ever :)
To not mentioned bugs :)
Regards,
Zoran
> I'm using 10g (10.1.0.3) on Linux and ASM and have
> run into bug 3458327
> which is also explained in Note: 264235.1. The gist
> of the problem is that
> ASM requires CSS to handle communications, but the
> way Oracle has setup up
> the startups, CSS is never available soon enough to
> allow the ASM instance
> to start. I've created a work around and wanted to
> share it, and see if
> anyone else has come up with a better way to handle
> the problem.
>
> Oracle suggests modifying the inittab file and then
> putting a "sleep" in
> dbora in the "hopes" that CSS will have initialized
> by the time dbora wakes
> up and tries to start the ASM instance. The problem
> is that I can't get the
> CSS to communicate with ASM unless I leave it where
> Oracle originally put
> it. Of course dbora then fails to properly start ASM
> and this must then be
> done manually after the boot. The following is
> Oracle's suggestion and my
> work around:
>
> Oracle's suggestion:
>
> 1 - put the entry in 'inittab' file about the
> init.cssd script before
> running the runlevel 3
>
> example of inittab:
>
> (...)
> s2:23:wait:/sbin/rc2 >/dev/msglog 2<>/dev/msglog
> s3:3:wait:/sbin/rc3 >/dev/msglog 2<>/dev/msglog
> s5:5:wait:/sbin/rc5 >/dev/msglog 2<>/dev/msglog
> (...)
> h1:3:respawn:/etc/init.d/init.cssd run >/dev/null
> 2>&1
>
> you need to change it in this way:
>
> (...)
> s2:23:wait:/sbin/rc2 >/dev/msglog 2<>/dev/msglog
> h1:3:respawn:/etc/init.d/init.cssd run >/dev/null
> 2>&1
> s3:3:wait:/sbin/rc3 >/dev/msglog 2<>/dev/msglog
> s5:5:wait:/sbin/rc5 >/dev/msglog 2<>/dev/msglog
> (...)
>
> 2 - amend the dbora script file to include a 'sleep
> 120' giving time to the
> daemon to start before the other instances try and
> start.
>
> My workaround (that works) is to background a script
> to start everything
> and allow the rest of the rc scripts to keep running
> and finish. This will
> then allow the init.cssd script to run as originally
> intended and the
> backgrounded script will presumably run after that.
> A more sophisticated
> approach would probably also check to make sure that
> ocssd.bin (CSS) was
> running before trying to start Oracle:
>
> Leave inittab alone and change dbora:
>
> 'start')
> echo "Starting up ORACLE..."
> #
> echo "sleep 120" > /tmp/dbstart
> echo "/bin/su - $ORA_OWNER -c
> \"$ORACLE_HOME/bin/lsnrctl
> start\"" >> /tmp/dbstart
> echo "/bin/su - $ORA_OWNER -c
> \"$ORACLE_HOME/bin/dbstart\"" >>
> /tmp/dbstart
> echo "touch /var/lock/subsys/dbora" >>
> /tmp/dbstart
> chmod 700 /tmp/dbstart
> nohup /tmp/dbstart > /dev/null 2>&1
> #
> echo "dbora is done."
> ;;
>
> Any thoughts?
>
>
>
> Andy Rivenes
> Email: arivenes_at_llnl.gov
>
> --
> http://www.freelists.org/webpage/oracle-l
>
-- http://www.freelists.org/webpage/oracle-lReceived on Wed Dec 01 2004 - 03:43:48 CST
 |
 |