FW: power hit stories... (fwd)

Dan Kohn (dan@teledesic.com)
Mon, 9 Sep 1996 08:40:19 -0700

[Great story. It sure is hard to engineer things correctly. - dan *]

>From: Hans-Werner Braun
>Sent: Sunday, September 08, 1996 8:15 PM
>To: Network
>Subject: RE: power hit stories... (fwd)
>Forwarded message:
>From owner-nanog@merit.edu Sun Sep 8 17:57:43 1996
>From: "Ry Jones (ECA)" <a-ryj@microsoft.com>
>To: "'nanog@merit.edu'" <nanog@merit.edu>,
> "'Bill Sommerfeld'"
> <sommerfeld@apollo.hp.com>
>Subject: RE: power hit stories...
>Date: Sun, 8 Sep 1996 17:47:21 -0700
>X-Mailer: Microsoft Exchange Server Internet Mail Connector Version
>MIME-Version: 1.0
>Content-Type: text/plain; charset="us-ascii"
>Content-Transfer-Encoding: 7bit
>At a company I used to work for, we had the badge reader/controller for
>the building in the security officer's office instead of the ops room.
>We had a failure (the Inauguration Day storm in the Pacific Northwest)
>that lasted longer than the UPS the security officer had put on his
>desktop. You guessed it... the building went into failsafe mode; only
>keys worked for access. None of the ops had keys and we were
>locked out of the building. Somehow we got a hold of the security guy's
>home number and he came in and we put in an extension cord. Pretty
>stupid, but not as cool as your story.
>Related, but not to power, one Monday (holiday) the building went into
>Day Mode, which was quite a bit less secure than Night Mode. Someone
>forgotten to enter the impending Holiday as a Night/Weekend mode day.
>The main doors, loading dock, and all internal doors were wide open all
>day until someone noticed none of them were locked.
>a-ryj@microsoft.com, rjones@wicker.com
>MSN WAN tools group
>>From: Bill Sommerfeld[SMTP:sommerfeld@apollo.hp.com]
>>Sent: Sunday, September 08, 1996 5:21 PM
>>To: nanog@merit.edu
>>Subject: power hit stories...
>>Reminds me of the power hit at MIT one summer Saturday in 1988. Ask
>>Jeff Schiller about it some time...
>>Short version:
>>MIT had just installed a 5ESS as its new PBX; they put in the usual
>>batteries for backup, and opted to not include a generator just for
>>the switch, because MIT had its own generator.
>>All the multi-line phone sets were ISDN sets, powered from the phone
>>closets. The phone closets supposedly had emergency power (from the
>>generator) and (small) backup batteries.
>>Anyhow, something bad happened somewhere in Cambridge (transformer
>>explosion or some such), with the result that most of town lost power.
>>For a long time.
>>One might expect the generators to come on.
>>Nope. They were off-line for maintainance.
>>Phys plant had brought in portable generators on trailers.
>>They didn't work either.
>>So, there were now three grades of power at MIT, and two of them
>>weren't working.
>> - normal power (from the local power grid, now dead)
>> - emergency power (from the generator, now dead)
>> - switch power (from the switch's batteries, still working)
>>Unfortunately, the main phys plant emergency number was on one of the
>>ISDN phones. As one might expect, that phone got a lot of calls after
>>the lights went out, and the phone went dead when its phone closet's
>>batteries went out..
>>Jeff was summoned to fix the switch so that the emergency number would
>>work again.. I happened to be nearby at the time and wound up getting
>>an impromptu tour of the switch.
>>We noticed, on arrival, that the lights in the nearby room housing the
>>modem pool were evidently on switch power.. which was clearly a wiring
>>Jeff tried to get into the switch room with his key card, but some
>>part of the card reader/electronic lock setup which controlled access
>>to the door to the switch room was not on switch power, and it
>>wouldn't let us in.
>>Our ears soon verified that the switch's burglar alarm, on the other
>>hand, *was* on switch power.
>>Once we shut up the alarm, we discovered that switch console terminal
>>was *not* on switch power...
>>Things went on like this for a while..
>>If there are multiple grades of power available, make *darn* sure that
>>nothing plugged into the highest grade depends on anything plugged
>>into a lower grade..
>> - Bill