Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 28 Aug 1996 21:24:33 -0400 (EDT)
From:      Brad Parker <brad@american.com>
To:        freebsd-stable@freebsd.org
Cc:        mbone@ISI.EDU
Subject:   corrupt cache with stable + adaptec 2940 + mbone?
Message-ID:  <199608290124.VAA19129@muir-woods.american.com>

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

I've been running FreeBSD stable (2.1.5-release, actually) for some
time now and every once and a while it freaks out and everything
starts dumping core.

Tonight it seemed to corrupt the disk cache (i.e. cached inodes) as
sed stopped working (it would just hang and eat cpu).  This was after
after emacs and gcc dumped core once, but them worked fine.  Sed had
been working fine all day.

I reinstalled sed from /usr/src and then it worked fine again. (I did
this assuming it would clear the in-core-inodes as we used to call them)

My problem seems correlated with problems on the mbone - today I got
hundreds of "Aug 27 13:30:54 muir-woods mrouted[52]: warning -
received packet shorter (576 bytes) than hdr+data length (20+1449)"
type messages on my console...  I'm running mrouted 3.8 and using this
machine as a multicast router, although no one was using the mbone
inside our net, to the best of my knowledge.

The last time I experienced this sort of bit rot was few weeks ago
when someone unleashed a few thousands routes...  Very similar result
- programs start dropping core.  I reboot and everthing is fine.

Anyone else seen this?  (I suspect a bug in the mbone code in
freebsd-2.1.5)

P120, 32mb, Adaptec 2940, SMC8416T ethernet card...

-brad




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608290124.VAA19129>