Comments on: Who’s In Charge Of The Time? http://davybrion.com/blog/2009/05/whos-in-charge-of-the-time/ Trying to walk that thin line between intelligence and ignorance Thu, 29 Jul 2010 21:01:23 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Oliver Hallam http://davybrion.com/blog/2009/05/whos-in-charge-of-the-time/comment-page-1/#comment-15059 Oliver Hallam Tue, 12 May 2009 17:19:27 +0000 http://davybrion.com/blog/?p=1356#comment-15059 We had a problem with a test for our XQuery date-time implementation that started failing when we implemented nightly test runs, but only on the nightly runs. The test checked that after adjusting the time zone of the current time it was still equal to itself. After spending some time checking the configuration of the build server to find out why that was the only place it was failing, it eventually turned out that the test was invalid between 23:00 and 24:00 due to a quirk in the specification. We had a problem with a test for our XQuery date-time implementation that started failing when we implemented nightly test runs, but only on the nightly runs.

The test checked that after adjusting the time zone of the current time it was still equal to itself.

After spending some time checking the configuration of the build server to find out why that was the only place it was failing, it eventually turned out that the test was invalid between 23:00 and 24:00 due to a quirk in the specification.

]]> By: Davy Brion http://davybrion.com/blog/2009/05/whos-in-charge-of-the-time/comment-page-1/#comment-14589 Davy Brion Thu, 07 May 2009 06:51:20 +0000 http://davybrion.com/blog/?p=1356#comment-14589 @Eric no idea how it's all configured... i just told the network admin about it and he said he'd look into it :) @Eric

no idea how it’s all configured… i just told the network admin about it and he said he’d look into it :)

]]> By: Eric Smith http://davybrion.com/blog/2009/05/whos-in-charge-of-the-time/comment-page-1/#comment-14538 Eric Smith Wed, 06 May 2009 20:33:11 +0000 http://davybrion.com/blog/?p=1356#comment-14538 "Normally, all of our servers’ system clocks are synchronized, but for some reason, the buildserver hadn’t had its system clock synchronized since the day before" Wow - so the build server managed to get out of whack by four minutes in a day? Am I understanding that correctly? Has an NTP service been installed and configured correctly on all machines involved? I do agree with you though that there should be a common service through-which all "current" date/times are acquired. It's then incumbent on the service to ensure that the date and time are always correct. “Normally, all of our servers’ system clocks are synchronized, but for some reason, the buildserver hadn’t had its system clock synchronized since the day before”

Wow – so the build server managed to get out of whack by four minutes in a day? Am I understanding that correctly? Has an NTP service been installed and configured correctly on all machines involved?

I do agree with you though that there should be a common service through-which all “current” date/times are acquired. It’s then incumbent on the service to ensure that the date and time are always correct.

]]>