On 4/21/14 12:26 PM, Buck Calabro wrote:
If you can afford to wait, see if you can get to a command line and
change your job priority to 2 and then WRKACTJOB, WRKSYSSTS to see
what's going on in there. If you can't afford to wait, force an IPL and
look at QHST to see if there's a clue as to what happened.

It had already been at least 15 minutes between onset and when I first screamed to the List for help, and close to another 15 minutes before I saw your message, and nothing had responded.

Some contexts were being updated in a Tomcat server at the time everything locked up. The history log shows a whole bunch of jobs (most of them Java jobs, but also some native jobs that our Tomcat contexts access) starting and ending between 11:46 and 11:53. Then, at approximately the same time I screamed for help, it shows a CPF0A9B (Probe failed for category QHTTP.) from job QYPSPFRCOL. Nothing else until the first message from the IPL at 12:28.

As I said earlier, I did a front-panel shutdown, with the pull-out control panel open, holding the recessed white button until the countdown on the screen got to zero. I see that the history log shows

CPI091D Previous ending abnormal, reason code 2.

of which which the helptext says
2 - The Power Down System (PWRDWNSYS) command did not complete.

No interrupted data base object level operations found.

Hmm: This is interesting: I see some messages in the general vein of
Job 059916/QUSER/QZRCSRVS ended on 04/21/14 at 12:29:04; .000 seconds used; end code 70

for jobs that had been active at the time of the front panel shutdown.

Nothing looks especially suspicious.


This thread ...


Return to Archive home page | Return to MIDRANGE.COM home page