Joel:
I worked a DC in downtown Chicago in the 70's and 80's and we were
supposedly 24X7 shop. We had power problems+ and we could not afford
a UPS in fact at the time we would have needed a HUGE UPS to get us
through power outages. I guess these were intermittent rather than
lengthy. Our biggest thorn in out side was a solid state paging
device as when it lost power you had to re-init it and re define the
PLPA page data set. Vsam at the time was a PITA as you couldn't
delete the pagespace unless it was already there so we ended up with
5 or 10 page data sets cataloged on the volume that was empty. (and
no noscratch was ignored most of the time).
Since we only used it for PLPA we had to IPL delete/define PLPA and
update parmlib and then re ipl with CLPA this became a weekly
occurrence until the big boss got tired of the extra outages and got
rid of solid state device (sorry do not remember the vendor).
I think we were all happy to see the monster go.
Ed
On Mar 14, 2013, at 9:55 AM, Joel C. Ewing wrote:
Ah yes, the not-so-good pre-UPS days when even a fraction-of-a-
second power interruption could take down an entire DP center. I
recall the utility workers describing at least one roasted squirrel
event and even one roasted large-bird event that affected our
transformer's HV feed, although more of the problems were weather-
related or human-related. One amazing thing was that in the 1980's
our utility power was much more stable, rarely above three or four
disruptive glitches a year. By 2000, noticeable utility power
glitches which reached our site had risen to several a month and
often were uncorrelated with local weather, but fortunately by then
the center was UPS-protected.
UPS-protecting all mainframe system hardware did significantly
reduce number of hardware failures we experienced. Even if you
have an environment where an abnormal shutdown and restart of the
DP center is is tolerable, an 8-hour or longer down time waiting on
parts might not be. These days, running a mainframe system without
adequate UPS protection should be considered false economy.
Joel C Ewing
On 03/14/2013 07:50 AM, John McKown wrote:
I think that "squirrel event" refers to a number of times that the
IBM-MAIN
listserv went down due to a suicidal squirrel throwing itself onto a
transformer, causing a massive power surge, which cause a
catastrophic
event in the computer center.
On Thu, Mar 14, 2013 at 7:45 AM, zMan <[email protected]> wrote:
On Thu, Mar 14, 2013 at 12:05 AM, Ed Finnell <[email protected]>
wrote:
Bon jour! We actually prosecuted a group of students who had
managed to
overlay the weekly backup tapes with garbage and hoped for a
'squirrel
event'.
They came perilously close...
I don't mind looking dumb: what's a "squirrel event"? Sounds like
something
Macy's would advertise...
--
zMan -- "I've got a mainframe and I'm not afraid to use it"
--
Joel C. Ewing, Bentonville, AR [email protected]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
