windows database service not shutting down database on reboot 2006-06-06 - By Paul Drake
On 6/6/06, Joe Armstrong-Champ <joseph.armstrong-champ@(protected)> wrote:
>
> We have about 10 windows servers, some w2k, some w2003 with 9i and 10g
> databases on them. When the server gets rebooted the database services
> don't get stopped and the instances crash. They always recover on
> startup and we have put a work-around in place but I'm curious about
> whether anyone else has experienced this problem. The following message
> appears in the oradim.log file when the problem happens:
>
> ORA-12638 (See ORA-12638.ora-code.com): Credential retrieval failed
>
> My guess is that a process which the database uses to authenticate
> itself has stopped before the oracle service and that is why the
> authentication fails.
>
> I have opened an sr but they are telling me to recreate the service
> using oradim locally, not remotely. I did this but when it didn't work
> they told me to apply the latest patchset and do it again. Since this is
> happening across different oracle and windows versions I have a feeling
> that it has nothing to do with the version.
>
> Has anyone else experienced this?
>
> Thanks.
> Joe
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
Joe,
Did you do any mucking about in the registry regarding
HKLM\SYSTEM\CurrentControlSet\Control\WaitToKillServiceTimeout
The default value is only 20 seconds (20000), which was likely fine for a
7.3.4 database using say 128 MB of physical memory and perhaps a single job
queue process.
A 10g R1 database instance using say 1.7 GB of physical memory may take
quite awhile to shutdown cleanly from a shutdown immediate. Perhaps the
following would help:
oradim -edit -sid mydb -shutmode a
or if you don't want to go the route of shutdown abort when the serivce is
stopped:
oradim -edit -sid mydb -timeout 600
This won't help in the case of a hard shutdown (power-off by BOFH, ups runs
out of juice, etc).
It also won't help with your current symptoms.
The re-create the serivce and apply the latest patchset are probably the
first 2 lines off of their script for this issue.
What credentials is the OS service running under? (localsystem, service
account with membership in the local OS ORA_DBA group, service account with
membership in the local OS ORA_MYDB_DBA group, etc)
Paul
On 6/6/06, <b class="gmail_sendername">Joe Armstrong-Champ</b> <<a href=
"mailto:joseph.armstrong-champ@(protected)">joseph.armstrong-champ@(protected)</a>
> wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote"
style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex;
padding-left: 1ex;">
We have about 10 windows servers, some w2k, some w2003 with 9i and 10g<br
>databases on them. When the server gets rebooted the database services<br>don't
get stopped and the instances crash. They always recover on<br>startup and we
have put a work-around in place but I'm curious about
<br>whether anyone else has experienced this problem. The following message<br
>appears in the oradim.log file when the problem happens:<br><br>ORA-12638 (See ORA-12638.ora-code.com):
Credential retrieval failed<br><br>My guess is that a process which the
database uses to authenticate
<br>itself has stopped before the oracle service and that is why the<br
>authentication fails.<br><br>I have opened an sr but they are telling me to
recreate the service<br>using oradim locally, not remotely. I did this but when
it didn't work
<br>they told me to apply the latest patchset and do it again. Since this is<br
>happening across different oracle and windows versions I have a feeling<br>that
it has nothing to do with the version.<br><br>Has anyone else experienced this?
<br><br>Thanks.<br>Joe<br>--<br><a href="http://www.freelists.org/webpage
/oracle-l">http://www.freelists.org/webpage/oracle-l</a><br><br><br></blockquote
></div><br>Joe,<br><br>Did you do any mucking about in the registry regarding
<br><br>HKLM\SYSTEM\CurrentControlSet\Control\WaitToKillServiceTimeout<br><br
>The default value is only 20 seconds (20000), which was likely fine for a 7.3.4
database using say 128 MB of physical memory and perhaps a single job queue
process.
<br><br>A 10g R1 database instance using say 1.7 GB of physical memory may take
quite awhile to shutdown cleanly from a shutdown immediate. Perhaps the
following would help:<br><br>oradim -edit -sid mydb -shutmode a<br><br>
or if you don't want to go the route of shutdown abort when the serivce is
stopped:<br><br>oradim -edit -sid mydb -timeout 600<br><br>This won't help in
the case of a hard shutdown (power-off by BOFH, ups runs out of juice, etc).
<br>It also won't help with your current symptoms.<br><br>The re-create the
serivce and apply the latest patchset are probably the first 2 lines off of
their script for this issue.<br><br>What credentials is the OS service running
under? (localsystem, service account with membership in the local OS ORA_DBA
group, service account with membership in the local OS ORA_MYDB_DBA group, etc)
<br><br>Paul<br>
|
|
