I'm having a similar problem, only, in my installation it appears that multi-recipient emails are being queued forever by postfix and then growing the dbmail-lmtpd process to an obscene size.
It is very bizarre, and actually gets worse (uses more resources) if I specify dbmail-smtp as the transport for the domains in question. Any tips on debugging this? The output from verbose logging of dbmail-lmtpd has not proven helpful, though it did fill up my root partition when I accidentally left it on overnight (d'oh!). Specific procedures for debugging dbmail that you guys use in-house would be great. I patch DBmail to work with our local schema and users database, but up until 2 weeks ago it was running like a charm. Now I see messages queued for 8, 9, 10, even 14 days, and I'll be damned if I can figure out why dbmail-lmtpd is dying. I throttled down MAX_CHILDREN to 2, MAX_CONNECTS to 2, and MAX_ERRORS to 1 for dbmail-lmtpd, just to get the resource consumption under control. It's all very bizarre. -- "To be suspended from the legal profession is the moral equivalent of being ostracized by child molesters." --Ian Rowan
