On-Call Hello Friday! And hello, therefore, to On-Call, The Register's regular column in which readers explain how they were sent out into user-land to do odd things and returned triumphant, frustrated or smugly satisfied.
This week, meet “Ben” who wrote to tell us that he once ran support for “a small mainframe manufacturer.” Among his jobs was tending the company's data centre and its own fleet of three machines.
“One summer,” Ben told us, “I got a series of support tickets about spontaneous reboots in the computer room.”
“Anyone who ran mainframes hates this type of problem. No log data, no record of power glitches on the UPS, in fact nothing except unhappy admins and irate users.”
Ben tried to sort things out, but couldn't. Meanwhile the frequency of outages “rocketed.” So he decided the only thing to do was to “sit in the corner of the computer room” and just see what was going on.
After half an hour of nothing at all happening Ben noticed a female colleague walking past in a loose dress, and mere seconds later the mainframe rebooted.
Ben conducted further investigation and learned that his colleague's dress was made of polyester. He quickly hypothesised that the combination of a whole lot of polyester, nylon carpet, plus dry, dry California air on a 105°F day might have been too much for a mainframe's anti-static protection.
“My electrostatic discharge engineer more or less called me a raving nutcase for suggesting this” Ben says, so he had to set up his own test. “Sure enough, even a low voltage spark crashed the system, but only in the area where the bus and tag cables entered.”
The fix was simple: a set of ferrite cores kept the stray jolts under control and Ben says he also arranged a payment so his colleague could “buy a wardrobe of cotton dresses (and underwear)” on the grounds that this was “better safe than sorry!”
Have you identified an odder fault than Ben's polyester problem? If so, write to share your story and it might be you appearing in this august space next week! ®