Need onstat Interpretation Help

Good Morning, IDS 11.50.FC5 O/S AIX 5.3 I have a recurring locking issue that I am trying to get to the bottom of. I have a script that will detect a lock held for > 15 seconds and among other things, grab an onstat -a and some session information for the holder of the lock as well as the waiter before killing the session holding the lock. During the course of chasing this all down in the onstat -a part, I get to a place that I am asking myself if what I am seeing is what should be there. Let me explain: 1. user 2 is waiting on a lock that user 1 owns. onstat -u and onstat -k confirm this. 2. user 1 is in a transaction and waiting on a condition (netnorm) onstat -u & onstat -g con confirm this. 3. Where I am getting confused is in the -g con output line in the waiter column. There is a different session id. Shouldn't that be my user 1 session? Side note, there are 58 waiters on this line. I know this is pretty cryptic w/o the benefit of actually looking at the onstat info but if you can answer my question. I would be thankful. Dan

onstat -g con shows thread IDs, not session IDs. Check if this gives somesense to your outputs. If not, you may need to share parts of it with us... Regards. On Fri, Feb 25, 2011 at 1:19 PM, DAN MUELLER <dan.mueller@trnswrks.com>wrote: > Good Morning, > > IDS 11.50.FC5 > O/S AIX 5.3 > > I have a recurring locking issue that I am trying to get to the bottom of. > I > have a script that will detect a lock held for > 15 seconds and among other > things, grab an onstat -a and some session information for the holder of > the > lock as well as the waiter before killing the session holding the lock. > During > the course of chasing this all down in the onstat -a part, I get to a place > that I am asking myself if what I am seeing is what should be there. Let me > explain: > > 1. user 2 is waiting on a lock that user 1 owns. onstat -u and onstat -k > confirm this. > > 2. user 1 is in a transaction and waiting on a condition (netnorm) onstat > -u & > onstat -g con confirm this.> > 3. Where I am getting confused is in the -g con output line in the waiter > column. There is a different session id. Shouldn't that be my user 1 > session? > Side note, there are 58 waiters on this line. > > I know this is pretty cryptic w/o the benefit of actually looking at the > onstat info but if you can answer my question. I would be thankful. > > Dan > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --0015174c3f22b833f2049d1b959c

DOH!!! Thank you for pointing out my early morning brain fart. Dan

Original Post: Good Morning, IDS 11.50.FC5 O/S AIX 5.3 I have a recurring locking issue that I am trying to get to the bottom of. I have a script that will detect a lock held for > 15 seconds and among other things, grab an onstat -a and some session information for the holder of the lock as well as the waiter before killing the session holding the lock. During the course of chasing this all down in the onstat -a part, I get to a place that I am asking myself if what I am seeing is what should be there. Let me explain: 1. user 2 is waiting on a lock that user 1 owns. onstat -u and onstat -k confirm this. 2. user 1 is in a transaction and waiting on a condition (netnorm) onstat -u & onstat -g con confirm this. 3. Where I am getting confused is in the -g con output line in the waiter column. There is a different session id. Shouldn't that be my user 1 session? Side note, there are 58 waiters on this line. I know this is pretty cryptic w/o the benefit of actually looking at the onstat info but if you can answer my question. I would be thankful. Dan Response: As already pointed out, the waiter output in the onstat -g con is a thread id. But what I think might be more valuable is what "netnorm" means. Which is that the session is waiting for data from the front-end client. Jacques Renaut IBM Informix Advanced Support APD Team