Time bobbles

[Olson, Arthur David (NIH/NCI) [E]]

Full marks for remote diagnostician Paul Eggert:

	Script started on Thu 29 Mar 2007 01:39:00 PM EDT
	lecserver$ who -b
	   .       system boot  Jan 22 14:40
	lecserver$ ls -l /usr/share/lib/zoneinfo/America/New_York
	-rw-r--r--   2 root     other       3519 Feb 12 17:44
/usr/share/lib/zoneinfo/America/New_York
	lecserver$ exit
	
	script done on Thu 29 Mar 2007 01:39:23 PM EDT

And, with a sendmail daemon restart, this (and subsequent) messages
should be correctly stamped.

The hidden perils of a long-running system...

				--ado

-----Original Message-----
From: Paul Eggert [mailto:eggert at cs.ucla.edu] 
Sent: Thursday, March 29, 2007 1:14 PM
To: tz at lecserver.nci.nih.gov
Subject: Re: Time bobbles

> Received: from lecserver.nci.nih.gov (localhost [127.0.0.1])
> 	by lecserver.nci.nih.gov (8.13.6/8.13.6) with ESMTP id
l2SDSIlX004500;
> 	Wed, 28 Mar 2007 08:28:18 -0500 (EST)
> Received: (from olsona at localhost)
> 	by lecserver.nci.nih.gov (8.13.6/8.13.6/Submit) id
l2SDS1lb004487;
> 	Wed, 28 Mar 2007 09:28:01 -0400 (EDT)
> ...
> Date: Wed, 28 Mar 2007 09:27:53 -0400
> From: "Olson, Arthur David \(NIH/NCI\) [E]" <olsona at dc37a.nci.nih.gov>
> 
> Word from a Washington, DC-area AOL subscriber is that their outgoing
> electronic mail is still being marked with "EST" time stamps that are
an
> hour behind wall clock time. Do folks know of other high-visibility
> bobbles of the change in US DST rules?

There was a bobble in your message headers (look for "EST" above);
does that count as high visibility?

Most likely your message's bobble was because the main 'sendmail'
daemon on lecserver.nci.nih.gov started running before its tz tables
got patched.  (The sendmail submit process on the same host knew the
right time zone.)

The tz database was fixed in August 2005.  Do you happen to know when
the patch was propagated to lecserver.nci.nih.gov, and when was its
sendmail daemon last restarted?...