On Thu, 2007-02-08 at 17:52 -0500, Sven Willenberger wrote:
> On Thu, 2007-02-08 at 17:37 -0500, Andrew Sullivan wrote:
> > On Thu, Feb 08, 2007 at 05:15:36PM -0500, Sven Willenberger wrote:
> > > prepareTableForCopy() function fire and the COPY command go. The table
> > > of 3 million rows finished copy from stdin fairly quickly but now the
> > > finishTableAfterCopy(115); analyze tablename has been "running" for some
> > > 40 minutes now; the only indexes on the thing are the primary key
> > > (two-column primary) and a single-column index (which "by hand" can be
> > > recreated in seconds). The cpu that this process has landed on is
> > > running at near 100% but a ktrace dumps nothing. Ideas?
> > 
> > Are there any ungranted locks?  I assume it's trying to add indexes
> > back, and it might have a conflict?
> > 
> > A
> > 
> Not sure what you mean by ungranted locks. The schemas were created by a
> regular user (i.e non-superuser) while slony is a superuser and should
> be able to restore the indexes. What should I look for to match your
> question?

As a followup, I went ahead and commented out the reindex line in the
function. After stopping and starting slony (to force a rollback and
re-attempt) it went through without a hitch. I did a quick analyze on
the table and a quick explain select to make sure the indexes were being
used and it all looked good. Something with the reindex on this
particular table is "bad" - it may be a Postgresql thing instead of a
slony thing here as when I tried to issue the reindex command as the
slony user it hung again.

Thanks for pointing me in the right direction at least :)

Sven

_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general

Reply via email to