[7967] in testers
Re: athena 10 /etc/athena/version
daemon@ATHENA.MIT.EDU (Camilla R Fox)
Mon Mar 23 17:23:55 2009
Message-Id: <200903232123.n2NLN4Wi029313@red-herring.mit.edu>
To: "andrew m. boardman" <amb@MIT.EDU>
cc: testers@MIT.EDU
In-Reply-To: Your message of "Mon, 23 Mar 2009 14:47:07 EDT."
<200903231847.n2NIl7Bh032047@pothole.mit.edu>
Date: Mon, 23 Mar 2009 17:23:04 -0400
From: Camilla R Fox <cfox@MIT.EDU>
> Are you looking at /etc/athena/version directly or via athinfo? While it
> seems like a reasonable change either way the latter was recently changed
> to not cat that file but instead run "echo `machtype -L` on `machtype -E`".
> (In the latter case, adding "installed at <date>" seems easy enough if
> there's a good place from which to pull a timestamp.)
I had in fact cared about the athinfo output; I liked the old athinfo
behavior of spewing the entire version file, because it's a good place
to start when you ask "is something amiss with this cluster machine".
(In particular, a cluster machine that's been reinstalled in the last
day is suspect for all sorts of reasons, as is one that shows evidence
of not taking updates.)
(I don't know what the other side of your design decision there, was;
for my uses, the cluttered output was easily piped through tail -1 when
not needed.)
In this case, I woke up to a buffer full of syslog spew which I tried to
make sense of before getting far enough through my mail to realise that
debathena machines had been intentionally installed in clusters.
> I think this has been addressed in the more recent round of mail, but if
> you're still seeing it please let us know.
Well, the debathena machines are noisy in a lot of ways I haven't
reported on; I think "what do you expect to get out of central logging?"
is a question that should be on the table, and what we do with syslog
output may follow accordingly.
-Camilla
