HP3000-L Archives

May 1996, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"<Elbert E Silbaugh>" <[log in to unmask]>
Reply To:
Date:
Wed, 8 May 1996 10:18:48 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (69 lines)
Good listeners:
 
After digesting the various comments on UPS, etc, I shall share some
thoughts.
 
A year ago we swapped our /70 for a /928RX with:
- 128mb memory (then)
- (3) 2gb disc drives
- 1.3kVA UPS
 
Other stuff too, but they are not plugged into the UPS.
 
For years, my sites have had jobs to bring up apps and supporting system
processes from scratch (i.e. system start) and to shutdown apps and
system processes, jobs, utilities, etc.
 
HP indicated that the UPS should keep the system alive for 15 minutes
or so. I decided before we went to production, that my curiosity get
satisfied. So I brought up the system to it's normal production
configuration (jobs, etc). Then I pulled the plug...
 
and waited...
 
and waited...
 
and waited...
 
about 2 hours and 40 minutes later the UPS died and the system shut down.
Then simulating a power restore in the building, I turned on the
DTCs and the UPS was plugged back into the building power. Much to
the surprise of my techie support guy, the system booted, and all my
apps, etc got started.
 
This was neat, I said. Now, how can I set it up so the system knows
when the UPS is about to die? Needless to say, I found out it couldn't.
So close, but yet so far!
 
So I asked HP back then, why not let the System Manager do some
system config parameters for the following 2 settings:
 
(A) The name of the user job stream file which should be launched
    to gracefully shut down the system in the event of an eminent
    power failure.
 
(B) The max estimated time it takes for that job to complete.
 
With these 2 simple parameters, the existing UPS Monitor process
that now gives a stupid SA, could instead launch the shutdown job
 
 
 
 
 
 
 
should it determine that power remaining in the UPS is nearing the
time required to run the shutdown job.
 
Obviously, this shutdown job could do other things such as CALL THE
ANALYST, etc.
 
They have the UPS talking to the CPU now, they have a trust monitor
process running now, all it takes is 2 simple parameters. If we want to
additionally have some SYSVARs, that is fine also. But with these
SYSVARs (some already mentioned in this thread), it would take some
user programs to make use of them.
 
Elbert

ATOM RSS1 RSS2