5.2-RELEASE: Possible bug in filesystem code?
pete at altadena.net
Sat Jan 17 06:29:51 PST 2004
I had observed many times over a long time two things:
1. The background fsck can easily cause a panic with a disk-intensive
process going on, and it has not seemed to depend on available space for me.
I have disabled background fsck in rc.conf on *all* my systems because of
2. The other source of a panic is if the background fsck detects a condition
which would cause the "preen" fsck to abort and give the "run fsck manually"
message at bootup, it (usually?) panics the system, usually with "freeing
freed inode". I can't think of how to handle this case any better; this
is a "hobson's choice" condition :-(
More information about the freebsd-current