Re: urgent: production database hangs on mount stage and won't open

Are the trace files generated when you tried to re-start the instance?? If so they may point to the actual problem.

I've had something that sounds a little bit similar on AIX with 8.1.7 and it turned out that the SGA's memory segment had never been deallocated at a UNIX Kernel level and so when a restart was tried the instance was hooking back into the old segment, due to name being the same, but not working properly. One of the UNIX admins at that site was ex-IBM development and he managed to kill it for me. However we were getting a trace file with memory allocation errors in it, so as I said at the top look for a trace file to point the way!

Kind Regards,

Fraser McCallum
MVP Oracle Administration
www.brainbench.com

"chao_ping" <chao_ping_at_163.com> wrote in message news:3d1021e5$1_at_usenetgateway.com...
> Hi, grurus: This is a production database 8.1.7.2 on aix 4.3.3, there
> are 17 instances running on the same box. I shutdown immediate the
> database and it hangs.After wait for one hour, i do a shutdown abort.
> When i startup again, it failed to open the database , just hangs at
> this stage like: Sun Jun 16 09:36:40 2002 SMON: enabling cache recovery
> SMON: enabling tx recovery And the sql*plus window hangs at : database
> mounted, but never showed open. From another shell prompt, i sqlplus
> internal and query the instance, it showed instance open in read write
> state! But from remote node via listener, client process showed 0134:
> oracle not available. I once met such kind of questions on sun solaris,
> it is because of solaris bug.But never heard of this kind on AIX. Oracle
> support suggest reboot the server, but as there is 16 other instance on
> it, it is not the solution. Can anyone help? Great thanks.
>
>
>
> --
> Posted via dBforums
> http://dbforums.com
Received on Wed Jun 19 2002 - 07:28:12 CDT