[24538] in Hotline Meeting
WATER in W20-575A
daemon@ATHENA.MIT.EDU (David Krikorian)
Fri Sep 30 06:31:32 1994
Date: Fri, 30 Sep 94 06:31:20 -0400
From: David Krikorian <dkk@MIT.EDU>
To: hotline@MIT.EDU, dcns-cluster@MIT.EDU, op@MIT.EDU, network@MIT.EDU,
star-maintainers@MIT.EDU
Cc: rcaileff@MIT.EDU
One of the air conditioning units in W20-575A was spilling water under
the floor. The problem was with the unit in the north-east part of
the room, ahead of you as you enter through the main door. The A/C
and water problems appear to be resolved. (Everything looked fine
when I left W20, but we should check again some time today.) Anyone
interested in details should read on.
The condensate from the A/C unit (I didn't think of checking for a
unit number) was not being pumped out, so the reservoir overflowed and
left an approximately 1/4-inch layer of water on the floor beneath it.
The water spread mostly under computerless parts of the room, soaking
network wiring and some extension cords. The water does *not* appear
to have spread at all to the SIPB portion of the room, or to the floor
under any of the DCNS servers.
I have no idea how long the water had been collecting, but at some
point tonight it leaked through to a student activity office on the
4th floor. Following up on complaints, some people went to W20-575A
at 1:45am today. They said they followed the procedure on the machine
room door, calling both x3-8400 and x3-2624, but got no response. (I
later listened to their message on x3-2624, and it was more like an
FYI message than an attempt to talk to anyone before entering. I
saved the message for reference.)
The repair people then entered W20-575A with a key, tripping the
alarm. I found out what was happening via SIPB members in the cluster
and the SIPB office who were being disturbed by the noise. The CPs
were called to disable the alarm, but didn't have the alarm
combination. A SIPB member finally disabled the alarm after checking
IDs to verify the authenticity of the repair people. The alarm was
alarming for about 10 minutes, from 1:50am until 2am.
The CAC and Physical Plant personnel dealt well with the situation
once the alarm was off. The main repair guy from Plant (the same one
who did most of the work on the E40-008 A/C unit when we had the
high-humidity alarm this summer) replaced a pump. The shift
supervisor, Glenn Wilder, was on the scene for a while, as were an
electrician, a plumber, and a handful of others at one time or
another. The CAC supervisor tonight, Derrick Barnes, was there, too,
with a couple of custodians who wet-vac'd up most of the water. At
4:20am, the last of them left and declared the situation resolved.
Many thanks to SIPB, PPL, CAC, and all those other acronyms for
handling the situation well.
There were no reports of network or server trouble. The only machine
that went down was ops-tmp-w20, which I shut down to give the workers
better access to the A/C unit, to related pipes under the floor, and
to the water. (It is back up now.)
