update stats high - server hangs. What table?
Posted in 2001
Gary Quiring on 7.31UC5, SCO OpenServer 5.0.5, saw the server lock up during a weekend update statistics high and asked how to tell which table was being processed. Responders suggested querying sysdistrib constructed dates, running oncheck for corruption, and running update statistics one table at a time with logging. Gary planned to use Art Kagel's dostats; the outcome was not reported.
Topics: General Discussion
I'm running 7.31UC5 on SCO OpenServer 5.0.5. When I do an update stats high on the weekend the server locks up solid. There's no errors in online.log or /var/adm/messages or syslog. Is there somewhere in the system tables that could tell me what table was being processed when the server hangs? I'm thinking I might have a corrupt record that hangs the server. Thanks Gary Quiring
The following query will tell you which tables were updated and which
were not based on the constructed date:
select systables.tabname,syscolumns.colname,sysdistrib.constructed,mode
from sysdistrib,systables,syscolumns
where systables.tabid > 99
and systables.tabid = syscolumns.tabid
and sysdistrib.tabid=systables.tabid
and sysdistrib.colno = syscolumns.colno
group by 1,2,3,4
order by tabname,colname;
A better solution might be to run updates stats in a shell script one
tables at a time with some logging messages to indicate where the
failure occurs.
Bill
In article <k54r7tcng58itle8kggatkfa5o8ant1t3q@4ax.com>,
Gary Quiring <gquiring@netlabs.net> wrote:
> I'm running 7.31UC5 on SCO OpenServer 5.0.5. When I do an update
> stats high on the weekend the server locks up solid. There's no
> errors in online.log or /var/adm/messages or syslog.
>
> Is there somewhere in the system tables that could tell me what table
> was being processed when the server hangs? I'm thinking I might have
> a corrupt record that hangs the server.
>
> Thanks
> Gary Quiring
>
>
Sent via Deja.com
http://www.deja.com/
If you think it's a corrupt record, you might want to try oncheck. I
agree with the other poster in that writing a logging shell script might
be the best solution in oncheck fails to find anything.
In article <k54r7tcng58itle8kggatkfa5o8ant1t3q@4ax.com>,
Gary Quiring <gquiring@netlabs.net> wrote:
> I'm running 7.31UC5 on SCO OpenServer 5.0.5. When I do an update
> stats high on the weekend the server locks up solid. There's no
> errors in online.log or /var/adm/messages or syslog.
>
> Is there somewhere in the system tables that could tell me what table
> was being processed when the server hangs? I'm thinking I might have
> a corrupt record that hangs the server.
>
> Thanks
> Gary Quiring
>
>
--
# unrm /
ksh: unrm: not found
# man cpio
Sent via Deja.com
http://www.deja.com/
On Mon, 05 Feb 2001 14:30:48 GMT, dareb@my-deja.com wrote:
>The following query will tell you which tables were updated and which
>were not based on the constructed date:
>
>select systables.tabname,syscolumns.colname,sysdistrib.constructed,mode
>from sysdistrib,systables,syscolumns
>where systables.tabid > 99
>and systables.tabid = syscolumns.tabid
>and sysdistrib.tabid=systables.tabid
>and sysdistrib.colno = syscolumns.colno
>group by 1,2,3,4
>order by tabname,colname;>
>A better solution might be to run updates stats in a shell script one
>tables at a time with some logging messages to indicate where the
>failure occurs.
>
>Bill
>
Thanks Bill,
I downloaded Art's dostats and will run it this weekend and see what
happens.
Gary Q