Trog wrote: > On Fri, 2004-02-20 at 11:49, Mike Brodbelt wrote: > >>Hi, >> >>Just a quick note to the list to point out that the recent changes >>appear not to have fixed the longstanding problem with clamav-milter's >>connections to clamd dying and then gradually growing the process list. >> > > > And the clamd logs say what?
Nothing useful:- Feb 20 10:37:28 castor clamav-milter[11294]: clamd / ClamAV version devel-20040219, clamav-milter version 0.67d Feb 20 10:57:32 castor clamd[12955]: stream: Worm.SomeFool FOUND Feb 20 11:05:19 castor clamav-milter[15370]: write failure to clamd There were no further clamd messages. > And the file(s) it was processing at the time are where? I'm only using it to scan mail, and the filter is set up such that when it dies, sendmail just bypasses it. The message in question will have been delivered. It's not immediately obvious to me how I can connect the log message:- Feb 20 11:12:31 clamav-milter[17358]: Timeout waiting for a child to die to a specific piece of mail. The thread timeout on clamd is set to 300 seconds, so I presume the offending mail would have arrived at around 11:07:31. There are 4 possible candidate messages that were timestamped at 11:07:30, all of which were spam or bounces. I have copies of the queue and data files for all these messages, and can happily provide copies if that would help. Mike. ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ Clamav-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/clamav-users
