Oh, my... I've experienced this today... I knew I should have updated DBMail...Just for anybody's interest: first it looked like the server was restarting (but it wasn't in fact) as it was unavailable for about 5 minutes. Then finally it started replying pings, I ssh'd to it and saw in log/messages that all mbuf clusters had ehxausted. Well, I won't go into details, but finally my maillog file increased by 85 MB in less than an hour, and during 43 minutes dbmail-imap4d was making children as fast as he could :-)
I made a command (well, yes - learning awk or smth else is in my list =)) ):grep "^Aug 25 16." maillog.0 | grep dbmail.imap | cut -d "[" -f 2 | cut -f 1 -d "]" | sort | uniq | wc -l
and got... 99731!!!!... %-) it's 38 processes in a second... I will update DBMail! Even right now probably...oh, great.. the FreeBSD port is old.. will have to `hack' the port system again...
[root]# ps auxww |grep dbmail-imapd |wc -l6812Matt, I've recently fixed a race condition in the preforking code that would allow process-leakage. Please try a snapshot and let me know if you hit this again.Thanks, will do. I haven't seen this problem before so I'm not sure how long until I seen it again, but I will surly let you know if I do.
-- Best regards, Danil V. Gerun.
smime.p7s
Description: S/MIME Cryptographic Signature
