Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbjade.BERKELEY.EDU Path: utzoo!linus!decvax!ittatc!dcdwest!sdcsvax!ucbvax!ucbjade!network From: network@ucbjade.BERKELEY.EDU (Berkeley Network) Newsgroups: net.rumor Subject: Re: Computer Horror Stories Message-ID: <366@ucbjade.BERKELEY.EDU> Date: Mon, 24-Feb-86 05:17:36 EST Article-I.D.: ucbjade.366 Posted: Mon Feb 24 05:17:36 1986 Date-Received: Wed, 26-Feb-86 06:34:34 EST References: <1215@brl-smoke.ARPA> <5027@alice.uUCp> Reply-To: mwm@ucbopal.UUCP (Mike (I'll be mellow when I'm dead) Meyer) Organization: Missionaria Phonibalonica Lines: 31 In article <5027@alice.uUCp> ark@alice.UucP (Andrew Koenig) writes: >To make a long story short, they started looking for any differences >between the customer site and their test lab. They finally found it: >the packs sent to customers had quality control stickers on them. >After a few weeks, some of the glue would work its way loose and find >its way into the heads... Gee, that sounds like the RA81's that DEC is shipping these days. The Hard Disk Assembly (HDA) (Head+Disk Assembly?) is an air-sealed unit, to keep the dust out of the heads (good idea, that!). The final stage of assembly is to glue the two halves of the box together, with disk/heads/etc inside, then evacuate the HDA. Unfortunately, the glue they used has a tendency to crumble inside the HDA. The things can be expected to crash inside the first year of service. Us? We've got 30 or so of the critters :-(. While I'm here, I might as well relate the tale of a VMS site that had secrataries doing backups. Early one morning, they did the backups, and managed to leave the system drive write protected. Since swap and the system logs were all on that, it didn't run very well - seems that processes would hang waiting for something to swap out so they could be loaded. After about 4 hours, performance got bad enough they went into the machine room to see what was going wrong. Besides the write-protected drive, they found 1/2 box of paper under the console, covered with messages of the form: ERROR: Can't write to SYS$ERRLOG. An attempt was made to put said error on SYS$ERRLOG, of course :-). Similar bugs used to crash IBM systems!