Hi all. Last Friday my MX choked and I am trying to determine why (so I can
defend qmail because I am sure it will get blamed). I am a big qmail fan, but
my boss is not. Not that he does _not_ like qmail, just that he prefers to go
with the old standby's like sendmail, or commercial products like Netscape MS,
which we use internally on our mail hub. But he did allow me to go with qmail
when we got our new MX server in.
So anyway, my MX is a Sun E250, dual 300MHz UltraSPARC II's, 256MB of RAM (not
the 1GB I reported in another mailing, I was confusing my MX and my mail hub),
running qmail. At that time, it was scanning incoming mail for viruses using
Scan4Virus+Mcafee's uvscan, and was referencing both the RBL and RSS.
I was working from home last Friday when my boss sent me an email
telling me that the MX had run out of swap (256MB physical swap) and needed to
be rebooted. So I have gone searching my logs for telltale information but I
have found next to nothing. I noticed that "status" messages in my smtpd log
were up around 17/40 around the time of the reboot, but that is nothing
phenominal I don't think. It has been stable ever since (and was stable for the
87 days before that, but I only added Scan4Virus and RSS checking in the past
week or two).
OT: I have seen the occasional 40/40 ... should I be increasing some parameter
here? Any wild guesses about how many connections I should/could allow on this
hardware (with virus scanning + checking RSS, DUL and RBL)? Actully, since
March 1st 00:00 I have hit 40/40 14 times, with all those occuring in 2
distinct blocks.
--
______________________________________________________
Mark Drummond|ICQ#19153754|mailto:[EMAIL PROTECTED]
Gang Warily|http://signals.rmc.ca/
Kingston Linux Users Group|http://signals.rmc.ca/klug/
