Re: distributed transaction hangs

> > >From the description of the problem it sounds as if the problem might
> > be related to the distributed locks taken in RBS segments. Do these
> > sessions that issue remote queries commit. If not, change them to do
> > so. At the bottom of the referenced article is a link to another
> > article that demostrates that remote queries are transactions and hold
> > RBS entries until the transaction is terminated (commit/rollback).

If it is related to the locks, then should I see at the remote db, that there are some locks which have v$lock.block = 1 ? If so, then it is not related to the locks, since I didn't see any locks with block = 1 at the remote db.

> The problem could be just poorly performing SQL, that is, the
> distributed SQL needs to be tuned.

no, the SQL it is waiting for is not more than a simple INSERT of one row..

> -- notice "<==" that remote v$session.process = local v$process.spid
thanks for this one! is it basically what I needed. Only in my case it was like this:
at local db:
v$process.spid = 134932
and at the remote db
v$session.process = 1048:134932

so it has put some "1048:" in front of it, what ever that means.. but still I can catch the remote session now.

the query from your FAQ, which uses the x$k2gte view also works with me! So thanks a lot, now I'm ready to make some further investigations when the problems appears again.. I'll be back here then I guess ;)

br.Andrei Received on Mon Jan 08 2007 - 13:55:11 CST