Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 30 Sep 2003 22:35:46 -0400 (EDT)
From:      Robert Watson <rwatson@freebsd.org>
To:        "Michael W. Oliver" <michael@gargantuan.com>
Cc:        freebsd-stable@freebsd.org
Subject:   Re: panics on 24 hour boundaries
Message-ID:  <Pine.NEB.3.96L.1030930223013.10236B-100000@fledge.watson.org>
In-Reply-To: <200309292126.36578.michael@gargantuan.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On Mon, 29 Sep 2003, Michael W. Oliver wrote:

> Hi stable, nice you see you again.  I was one of those guys who was
> seeing constand panics on 24 hour boundaries but couldn't provide a
> backtrace due to the ar device not taking a dump.  I installed a
> dedicated drive just to take the dump, and then didn't have a panic for
> a couple weeks.  Now, I am back with, and I have traces to share. 
> 
> The first two, from 2003-09-27 and 2003-09-28 are from an older
> kernel...  the same one that was working great for a couple weeks.  Due
> to some severe storms around here, I shut the maching down voluntarily,
> and once rebooted, it started the 24 hour panic crap again.
> 
> So, since I didn't have anything much to lose, I cvsuped my src and
> rebuilt and installed world and kernel on the evening of 2003-09-28.  Lo
> and behold, I had another 24 hour panic.  The trace for this one is
> phoenix_2003-09-29.txt.  The full uname from this latest panic is: 
> 
> FreeBSD phoenix.gargantuan.com 4.9-PRERELEASE FreeBSD 4.9-PRERELEASE #1: 
> Sun Sep 28 18:16:30 EDT 2003
> root@phoenix.gargantuan.com:/usr/obj/usr/src/sys/PHOENIX i386
> 
> If this information isn't sufficient, please let me know what else I can
> do to help troubleshoot this problem.  Thanks in advance for your help,
> it is much appreciated! 

Initial reactions: panics on 24 hour boundaries are, in my experience,
often associated with the daily event.  Once a day, the daily scripts run
find several times on your file systems, causing every file and directory
to be inspected for changes in setuid scripts, etc.  This can trigger
certain classes of race conditions and resource limits that you might
otherwise not hit in normal operation -- and conviently, they run 24 hours
apart :-).  To try and confirm this suspicion, it would be interesting to
know what time of day exactly the panics take place, and whether you can
reproduce the panic by manually running the daily or security script.

Also, out of curiosity: since you're experiencing crashes, I assume fsck
has run on all your file systems.  If not, you might want to boot to
single user and run fsck on each file system manually to make sure there's
no on-disk corruption of UFS meta-data.

Robert N M Watson             FreeBSD Core Team, TrustedBSD Projects
robert@fledge.watson.org      Network Associates Laboratories



  [NON-Text Body part not included]




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.NEB.3.96L.1030930223013.10236B-100000>