Ok, I know, this sounds quite controversial, and it's probably not even
qmail's fault, but I'm curious if anyone else has had the same sort of
problem.
A user just sent 2 subsequent mails to a mailing list, both including quite
large attachments (looking in ~user/list/archive, the files are 1,2Mb and
730kb respectively) to 21 users. This is quite heavy on a 128kbit line, esp.
since concurrencyremote is 20, so I expect that each qmail-remote takes its
share of the bandwidth, leaving 128/20 kbit per delivery.
So, naturally, I'll get a few defferals due to dead connections, but
eventually all the mail should get sent, so I don't worry about that. What I
do worry about is that I can't login because bash can't fork (due to lack of
memory,) that the load peaks around 2.40 and averages at 1.20 and that the
TCP/IP stack is dead. (I can't get a ping reply from our local network and all
connections seem dead.)
What could be the cause of this? Could it be qmail or just an instable linux
kernel?
Thanks for any comments.
System information:
PPro 180MHz
64Mb RAM
100Mb Swap
3Com 3c905B (Running in 100Mbit to the router)
Slackware 4.0
Linux 2.2.12 (egcs-1.1.2)
Using syslogd for logging (I know, I know.. I need to change. How about
syslog-ng btw?)
On the same server, I also run
proftpd
apache
bind
but it's on quite a small scale, and shouldn't matter much.
Thanks, Henrik.
