RE: Syslog gone?
Chris wrote: > > I can't answer your questions, but that -m 16 has me worried. > > At around > > 40-60 megs per child, that is using up a LOT of ram. Unless > > you have a few > > gigs spare on this machine I'd stick with the default of 5 > > children until > > you have things up and running moderately well, then > > experiment with more if > > necessary. > LOL, thats the first thing that hit me as well! Sixteen!! > Hope they got some big iron! Remember, -m = "--max-children", not the number of permantly preforked children. It's just a ceiling to protect the system from overload (4G memory). I have 4 preforked children now, and 20 similarly-sized Apache processes; none of which is really out of the ordinary. - Mark
RE: Syslog gone?
Title: RE: Syslog gone? > -Original Message- > From: Loren Wilton [mailto:[EMAIL PROTECTED]] > Sent: Friday, October 06, 2006 4:15 AM > To: users@spamassassin.apache.org > Subject: Re: Syslog gone? > > > > /usr/local/bin/spamd -D -x -q -d -m 16 -r > /var/run/spamd.pid -i 127.0.0.1 > > -A 127.0.0.1 -u spamd -s mail.crit > > I can't answer your questions, but that -m 16 has me worried. > At around > 40-60 megs per child, that is using up a LOT of ram. Unless > you have a few > gigs spare on this machine I'd stick with the default of 5 > children until > you have things up and running moderately well, then > experiment with more if > necessary. LOL, thats the first thing that hit me as well! Sixteen!! Hope they got some big iron! --Chris
Re: Syslog gone?
Well, I got my syslog back. Apparently, 3.1.6 no longer accepts facility in the form of "mail.crit" (facility + severity; 2.55 did). But using a single word facility like "debug" or some such works fine. Yep. Facility coding scheme changed about three releases back as best I recall. You might find 'info' more appropriate than debug. Loren
RE: Syslog gone?
> -Original Message- > From: Loren Wilton [mailto:[EMAIL PROTECTED] > Sent: vrijdag 6 oktober 2006 10:16 > To: users@spamassassin.apache.org > Subject: Re: Syslog gone? > > > Well, I got my syslog back. Apparently, 3.1.6 no longer accepts facility in the form of "mail.crit" (facility + severity; 2.55 did). But using a single word facility like "debug" or some such works fine. - Mark
Re: Syslog gone?
/usr/local/bin/spamd -D -x -q -d -m 16 -r /var/run/spamd.pid -i 127.0.0.1 -A 127.0.0.1 -u spamd -s mail.crit I can't answer your questions, but that -m 16 has me worried. At around 40-60 megs per child, that is using up a LOT of ram. Unless you have a few gigs spare on this machine I'd stick with the default of 5 children until you have things up and running moderately well, then experiment with more if necessary. Loren
Syslog gone?
Ok, nearly there. :) I start spamd 3.1.6 as follows: /usr/local/bin/spamd -D -x -q -d -m 16 -r /var/run/spamd.pid -i 127.0.0.1 -A 127.0.0.1 -u spamd -s mail.crit I get: [11581] dbg: logger: adding facilities: all [11581] dbg: logger: logging level is DBG [11581] dbg: logger: successfully opened file mail.crit [11581] dbg: logger: successfully added file method [11581] dbg: spamd: creating INET socket: [11581] dbg: spamd: Listen: 128 "successfully opened file mail.crit"?? Say what? :) I specified the syslog facility "mail.crit", not a file! Far as I can tell, -s is still used to set the syslog facility. How do I get SA to log to syslog again? These are more worrisome, though: Fri Oct 6 09:04:18 2006 [11583] dbg: prefork: new lowest idle kid: 11585 Fri Oct 6 09:04:18 2006 [11583] dbg: prefork: child reports idle Fri Oct 6 09:04:18 2006 [11583] info: prefork: child states: II Fri Oct 6 09:04:18 2006 [11585] dbg: prefork: sysread(7) not ready, wait max 300 secs Fri Oct 6 09:04:28 2006 [11586] dbg: prefork: periodic ping from spamd parent Fri Oct 6 09:04:28 2006 [11586] dbg: prefork: sysread(8) not ready, wait max 300 secs Fri Oct 6 09:04:28 2006 [11585] dbg: prefork: periodic ping from spamd parent Fri Oct 6 09:04:28 2006 [11585] dbg: prefork: sysread(7) not ready, wait max 300 secs That does not look good. What does it mean, and how can I solve it? Thanks, - Mark
