Subject: | |
From: | |
Reply To: | |
Date: | Sun, 5 May 1996 08:35:17 EST |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Speaking of power-fail recovery, we use a 10KvA UPS under
what was a 3.4 KvA load which could be depended on to give
us about 40/45 minutes. Unfortunately, a few more devices
(including a DEC Alpha and about 6 assorted LM file servers)
were added to the circuitry just before our Engineering
Department decided to extend independent power generation
testing from 45 minutes to 1 hour (followed by switching
back to mains, and continuing to run the diesel generator
against a 750KvA load bank for another 4 hours).
Needless to say, the computer rooms with UPS don't get their
juice from the diesel generator. With UPS it was decided we
were 'covered' and could survive. Mistake #1 -- with electricity,
calculations have to be accurate, comprehensive, and frequently
reviewed.
Anyway, to make a long story short, we exhausted our batteries
both UPS and CPU, and had to do a RS at last power test.
Unfortunately, a crucial switch in one of the many (UB)
network boxes failed and it was about an hour before the
DTCs could be brought back up. As a result, the spooler
seemed to go into a permanent sleep and nothing we did could
get access to the system printer (or any other printer).
DTCs are controlled by HPOV, so we tried resetting them,
then power-cycling them ... no dice.
Another RS fixed everything, but ...
Was there anything else that could have been done? Maybe just
a simple network stop and start?
Ron
[log in to unmask]
|
|
|