Michael Dankov wrote: > Hi! > > Look, two minutes before SelfCheck there is milter thread 8880 started, and > I do not see when it finishes. Is something wrong with it?
I have no log entry of that one ever finishing. > At the moment I have no ideas what is going wrong in your case, following is > my clamav.conf, try to change your to be as close to it as possible, and > recheck if you had installed libclamav after last rebuild. > > === clamav.conf === > LogFile /tmp/clamd.log > LogFileMaxSize 2M > LogTime > LogSyslog > PidFile /var/run/clamd.pid > LocalSocket /var/run/clamav/clamd > FixStaleSocket > StreamSaveToDisk > ThreadTimeout 7200 > MaxDirectoryRecursion 15 > User clamav > ScanMail > === clamav.conf === This is my clamav.conf:- ================================ LocalSocket /var/run/clamd.ctl ScanMail ScanArchive StreamSaveToDisk StreamMaxLength 35M ArchiveMaxRecursion 5 ArchiveMaxFiles 1000 ArchiveMaxFileSize 15M ThreadTimeout 180 MaxThreads 5 MaxConnectionQueueLength 15 LogSyslog PidFile /var/run/clamd.pid DataDirectory /var/lib/clamav/ SelfCheck 3600 =============================== I'll play with your options and see if they make any difference. > It may be helpful if you try to run clamdscan on a file when milter > begin to block and look if clamdscan hangs too. Don't forget that clamdscan > can check only files accessible by user running clamd, not clamdscan. That appears to work - it fell over again this morning, and clamdscan was still seemingly OK. > MB>It then accumulated processes until I notice, and kill and restart it. > > Do you need to restart both clamav-milter and clamdscan, or only one of > them? I restart both clamd and clamav-milter. > MB>Something appears to have improved recently, beacuse with the 09122003 > MB>CVS, although I still get the huge max-children numbers in the logs, I > MB>don't actually seem to have hundreds of processes any more. > > I had switched that limit off for clamav-milter because at some point with > 0.60 it seemed it was not working properly. Yes - I think it is working better now, though still not 100% > MB>So, I think that the self-check is actually where it starts to go wrong, > MB>and that clamd falls over somehow, which causes the milter to block and > MB>time-out, and also means that it doesn't always exit cleanly (leading to > MB>the not NULL errors). > > At the moment I do not see any way for self-check to break things... > > May be, switching to process based scanning would help? Try UseProcesses > option in clamav.conf. Will have a play, and let you know. Mike. P.S. This is a dual processor box - any remote possibility of that leading to a race condition somewhere? I'd have thought not, but figured mentioning it can't do any harm. ------------------------------------------------------- This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click _______________________________________________ Clamav-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/clamav-users
