Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: Database won't mount, but no errors reported ???
As everyone else has said, this is a known issue on Solaris.
FYI... When Oracle hangs at startup, a good trick to see more informational error messages is to turn on trace at the server level with init.ora parm sql_trace=true. Not sure what it would say, but I know 1 time this came in handy when we hit the max open files set at the kernel level.
Just do not leave this parm in the init.ora, it generates a billion trace files.
Note: Has anyone ever heard of Windoze machines being up for over 7 months ? We schedule nightly reboots on most of our Windoze servers and thankfully I don't work on those. We , of course , run Oracle on a real OS.
"Mike Heden" <mheden_at_bigfoot.com> wrote in message
news:53b3de4.0301161226.1a864131_at_posting.google.com...
> A system running Oracle 8.0.6 under Solaris failed overnight.
> Attempts to restart it just hang. If I do a startup nomount, it
> reaches a 'started' state OK and server manager comes back with its
> prompt. On the other hand if I do a startup mount it hangs. There's
> nothing in the alert log that indicates any sort of problem - the last
> entry is 'alter database mount'. I've also left it a long, long time
> just in case it was taking its time loading, but that has made no
> difference. Once it's in this hung state, not even control-C gets a
> response - the telnet session to the server appears dead.. There are
> no dumps and there's nothing in dmesg to indicate a more fundamental
> problem.
>
> I've tried restoring from a full hot backup taken the previous night,
> but the results are exactly the same.
>
> Fortunately this is an inactive legacy system that is only used
> occasionally to provide access to old, virtually-static, data. My
> next step is probably going to be a restore from the last cold backup.
>
> Am I missing something here? I haven't seen Oracle 'fail silently' in
> this way before, so I'm inclined to think it's not a database problem.
>
> Any ideas?
>
> Thanks,
>
> Mike
Received on Sat Jan 18 2003 - 00:52:56 CST
![]() |
![]() |