[1897] in athena10
Re: Linerva outage today
daemon@ATHENA.MIT.EDU (Tim Abbott)
Thu Mar 26 15:41:27 2009
Date: Thu, 26 Mar 2009 15:40:28 -0400 (EDT)
From: Tim Abbott <tabbott@MIT.EDU>
To: Evan Broder <broder@mit.edu>
cc: linerva@mit.edu
In-Reply-To: <49CBD833.6020400@mit.edu>
Message-ID: <alpine.DEB.1.10.0903261538470.7289@vinegar-pot.mit.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
We should probably set up a permanent linerva test machine on XVM so that
we are not tempted to do debugging as root on the production system in the
future.
-Tim Abbott
On Thu, 26 Mar 2009, Evan Broder wrote:
> Linerva went down today at around 3:00 today. We were attempting to
> trace the "pag bug" that sometimes results in users being unable to
> access their homedirs. It seems that killing an ltrace process that was
> tracing a test ssh daemon also tried to kill the init process, which
> resulted in a kernel panic.
>
> We plan to do all future debugging of this bug on other machines so as
> to avoid future problems. We'll be attempting to track down both the
> "pag bug" and the ltrace bug and report them upstream.
>
> We're very sorry for the inconvenience this caused.
>
> - Evan
> for the Linerva maintainers
>
