[20207] in Athena Bugs
Re: LIDS Athena Cluster
daemon@ATHENA.MIT.EDU (Petr M. Swedock)
Fri Mar 1 07:56:50 2002
Date: Fri, 1 Mar 2002 07:56:49 -0500 (EST)
Message-Id: <200203011256.HAA11424@nerd-xing.mit.edu>
From: "Petr M. Swedock" <petr@MIT.EDU>
To: zacheiss@mit.edu
CC: zacheiss@mit.edu, bugs@mit.edu
In-reply-to: <200203010910.EAA04355@brad-majors.mit.edu> (message from Garry
Zacheiss on Fri, 01 Mar 2002 04:10:31 -0500)
: cc: zacheiss@MIT.EDU, bugs@MIT.EDU
: Date: Fri, 01 Mar 2002 04:10:31 -0500
: From: Garry Zacheiss <zacheiss@MIT.EDU>
:
: >> Yes, I see. I'm at a loss to explain the multiple, repeated, 'time
: >> sync' errors in /var/adm/messages... I've seen them in a bunch (haven't
: >> looked at all) of the machines in my cluster.
:
: Could you tell us what type of machine and what subnet you're seeing
: it on, as well as the exact text of the error from /var/adm/messages?
: There have been some problems with machines doing ntp sync all through
: the 9.0 release, although I thought they were mostly cleaned up.
Sure, I'll compile a small list, with examples, sometime today.
: I wouldn't worry about it too much, since even if ntp isn't time
: syncing, /etc/athena/gettime is run from cron twice daily to set the
: time, and AFS will also reset the machine's time if it gets too far out
: of sync. If there are still remaining NTP problems, it's something we
: should look into, but I wouldn't expect machines to be active losing
: because they're more than 5 minutes off from kerberos's clock.
We seem to avoid losing, which is good, but I worry about the margins
=-)
Thanks for your help.
Peace,
Petr
