Date: Tue, 17 Oct 1995 14:35:57 -0400 (EDT) From: Brian Tao <taob@io.org> To: FREEBSD-HACKERS-L <freebsd-hackers@freebsd.org> Subject: Machine lockup, IRC and FTP server Message-ID: <Pine.BSF.3.91.951017141903.207R-100000@flinch.io.org>
next in thread | raw e-mail | index | archive | help
I'm starting to see total machine lockup on my IRC/FTP server
(2.1.0-950928-STABLE, 486DX2/66, irc 2.8.21, wu-ftpd 2.4). The only
pattern I can deduce so far is that it happens after about 2 days of
uptime. During those two days, the machine runs beautifully. With 64
megs, it hardly touches swap even with a fully loaded EFnet IRC server
and 50-100 FTP connections. Then, out of the blue, it just freezes
up (always when I'm away from the office too :-/).
I built a new kernel with NMBCLUSTERS=1024, but that didn't seem
to help much. No warnings about running out of mbufs are sent to
syslog. In fact, nothing at all in syslog would indicate something
went wrong. The machine just locks up, no console switching, no
network, no kernel panic, nothing.
I've monitored the output from top, pstat -T, pstat -s, netstat -m
and netstat 1. The only abnormality is the high number of collisions
on our Ethernet (15% to 75%, but that's another story). I don't see
any wild fluctuation in the numbers that would cause the machine to
hang like this.
Any idea what I can do to tweak the kernel, or other stats I
should be monitoring? Is there a way to find out the NMBCLUSTERS for
a given kernel (assuming the kernel config is no longer available)?
--
Brian Tao <taob@io.org>
System Administrator, Internex Online Inc.
"Though this be madness, yet there is method in't"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.951017141903.207R-100000>