Re: aborting long transaction

> I am running ODS 7.23 on HP O/S 10.20. A few users have received some error messages. I looked at the log file and then added 14 more logs. But they got the same errors. I checked onstat -l and they were not full.> > What else I should look at? Here is the log file info. > > > 10:22:58 Checkpoint Completed: duration was 3 seconds. > 10:25:49 Logical Log 291 Complete. > 10:25:51 Checkpoint Completed: duration was 1 seconds. > 10:25:51 Aborting Long Transaction: tx 0xc26424a0 username: marlynr uid: 112 > 10:25:54 Process exited with return code 133: /bin/sh /bin/sh -c /usr/informix/> etc/log_full.sh 2 23 "Logical Log 291 Complete." "Logical Log 291 Complete." > 10:31:05 Checkpoint Completed: duration was 2 seconds. Not knowing your level of experience, I'm going to ask a couple of questions that you may have already considered. After you added the logs, did you do a Level 0 archive? If not, the new logs won't be used, thus you have the same amount of log space as before you added logs. I believe they show a status of "A" (or maybe "N") when you do an onstat -l if they are not yet eligible for use. Next, what are the values for LTXHWM and LTXEHWM? If you're not sure what these do, check the Admin Guide. If the logs are in fact available for use, and the LTXHWM/LTXEHWM values are reasonable, check with user marlynr to see what they were doing at the time of rollback. They may be trying to insert 500,000 records in one transaction, or they may be starting a transaction and then walking away for a couple of hours without doing a COMMIT or ROLLBACK. Mark Collins mcollins@us.dhl.com Dilbert is a documentary.