Paul J Stevens wrote: > Paul J Stevens wrote: >>Thomas Mueller wrote:
>>>I had as many connections as running processes (a lot below >>>MAXCHILDREN). The problem was: no one could connect any more. Looks like >>>a problem with starting new processes - every running process handled a >>>connection yet, no new were started so no user could connect. >> >> >>Thomas, >> >>I'd be very interested in your logs, especially the logs from the pool.c code. > > The main issue for me here is that I'm unable to reproduce this condition in > my > own setups. That makes tracking this down an academic excercise at best for > me, > without the confidence inspired by thorough testing, which I prefer. It simply > takes more time than simply tracking down a well documented bug in the code. I have to run dbmail at loglevel 1 or my syslog grows too fast, so there is not much I can provide, sorry. Uh - very, very strange. I did: cat /var/log/syslog |grep dbmail |grep pool.c > /tmp/dbmail.log The first lines are: Aug 18 15:41:04 geht-schon dbmail/pop3d[16031]: pool.c,manage_stop_children: General stop requested. Killing children.. Aug 18 15:41:04 geht-schon dbmail/pop3d[16033]: pool.c,child_unregister: child [16033] unregistered Aug 18 15:41:04 geht-schon dbmail/pop3d[16036]: pool.c,child_unregister: child [16036] unregistered Aug 18 15:41:06 geht-schon dbmail/lmtpd[16047]: pool.c,manage_stop_children: General stop requested. Killing children.. Aug 18 15:41:06 geht-schon dbmail/lmtpd[16049]: pool.c,child_unregister: child [16049] unregistered Aug 18 15:41:06 geht-schon dbmail/lmtpd[16052]: pool.c,child_unregister: child [16052] unregistered Normally I get regularly: Aug 18 23:24:16 geht-schon dbmail/imap4d[23710]: pool.c,child_register: register child [23710] Aug 18 23:24:17 geht-schon dbmail/imap4d[23710]: pool.c,child_register: initializing child_state [23710] using slot [4] Aug 18 23:24:17 geht-schon dbmail/imap4d[8703]: pool.c,manage_spare_children: children [14/200], spares [4 (4 - 8)] Aug 18 23:26:26 geht-schon dbmail/imap4d[24136]: pool.c,child_register: register child [24136] ... The syslog starts at 6:30 - 9 hours without any pool.c loglines! I hope that helps a bit. Thomas