>Folks,
>
>Sorry I should have provided a few more details initially.
>
>Here is the squence of events:
>
>03:00
>1. We shut down the engine with onmode -ky (probably no users present)
>2. The RS6000 box is rebooted.
>3. We restart the engine with oninit
>
>It during step 3 that the online instance hangs. Below are the lines
>from the online.log file.
>
> 02:05:40 Checkpoint Completed: duration was 0 seconds.
> 03:03:16 INFORMIX-OnLine Stopped.
> 03:13:51 VP pid=13632 priority fixed at 65, former = 114
>
> Sat Aug 23 03:13:58 1997
>
> 03:13:58 Event alarms enabled. ALARMPROG = '/usr/informix/log_full.sh'
> 03:13:58 VP pid=21060 priority fixed at 65, former = 65
> 03:13:58 VP pid=21318 priority fixed at 65, former = 65
>
>This is where is sits until I kill the VPs. If you do a onstat - the
>state will indicat Initializing. After the VPs are killed I issue a oninit
>and the engine starts up okay.
>
At the point of the hang, it might help to do the following:
1) get several onstat -g ath -- try to notice which threads remain in a
running state.
2) If it appears that one of the threads is remaining in a running state
within all of the onstats, then we are probably looking at a run-away
thread. Try to get several stack traces from that thread. This can be
done by issueing an kill -7 <unix process id of vp that appears to be
hung>. The processid can be done by taking the output from an onstat -g
glo and the vp number from the onstat -g ath.
Contact support and give them this information. It can be used to help
resolve the problem.
Madison Pruet