Subject says most of it.

We received pages on Sunday that the 3000 was down.  The IS Director came in
and found the system sitting at a boot prompt.  With nothing else to go on,
he performed a BO PRI, and a START NORECOVERY.  The system came back up
without any problems, and all processes were restarted.  On Monday, he
assigned me the task of determining the cause for the apparent system
self-bounce.  I didn't make a point about getting a DUMP.  I searched the
system log files and found nothing.  I determined what was running on the
system, and found that there was no unusual activity.  In essence, nothing
looked wrong.  Our UPS seems to be running fine, and no other equipment on
the same circuit was down or seems to have been effected.

This morning I come in to find a portion of the console log printout on my
desk.  He was in last night to restart the system after receiving pages that
we were down again.  Because I had not made the point clearly enough on
Monday that "without a DUMP" we would not be able to do much, unfortunately
we didn't  get taken last night either.  Our hardware support is through ICS
and I've already spoken to my tech there about getting their version of
predictive support running to check our hardware logs.  In the interim, we
have predictive support on our box and am thinking about jumping into a
manual or two to see how to check the hardware logs manually.

We are on a 939020KS, running 6.5 with Summit's patchset I2 (roughly
powerpatch 2, plus some).  On Saturday, we had both our ATM Network
interfaces, and Front Office User interface applications running (Realtime+
and Spectrum).  It was early evening, and otherwise, no real activity was
taking place.  (We have a weekend processing derth, for the time being.)
Last night, however, the Front Office application (Spectrum) was down, some
nightly reporting was taking place and our full system backup (Backup/iX)
had just finished processing to tape and was using ftp.arpa.sys to send the
SYSLIST to one of our W2K file servers.  In both situations our LAN was up,
JINETD was running, JobPak was processing $STDLISTs, Maestro was running,
and our ODBC background job was active.

The differences seem more glaring to me than the similarities.  So, without
anything else to go on I'm tending to think we have a hardware problem, but
I can't tell without a dump.  So, here comes the long shot....anyone got any
ideas about what I can check to determine what is causing these mysterious
problems?

Help,
Rs~

Russ Smith
Systems Analyst, CalState9 CU, Concord CA, rsmith @ calstate9.comm
Programmer/Analyst, Problem Solved, Vacaville CA, rsmith @ cu-help.comm
3000L lurker and troublemaker, work @ rsmith.orgg

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *