On a stock RedHat 9.0 box (3 boxen load balanced) with updated kernel (2.4.20-20.9smp), I have stability problems with clamd.
I'm using sendmail -> clamav-milter -> clamd. Our mail servers accept about 50K mail per day (each box), of which about 35K gets rejected by spamassassin before it ever reaches clamav. It seems that clamd runs between 1-3 hours and then dies with a segfault. I increased logging to its max (LogVerbose, LogTime, and LogClean) but didn't see any segfaults in the couple of hours I let it run. I stopped those and reran it in debug && foregound mode for another couple hours and didn't see a segfault there either. While this was going on, I was speaking to a coder friend of mine. His first question was "is this linked to pthread?" Yes it is, and this is a RH9.0 box with the NPTL patch applied. His gut feeling was that pthreads were causing a problem and had seen these exact types of intermittent problems. As an attempt to stabilize things, I added export LD_ASSUME_KERNEL=2.4.2 to the init script. I disabled foreground mode leaving on LogVerbose, LogTime, and LogClean and restarted the daemon. It's been 2 hours now and all three boxen have stayed up that entire time, so it may have made a difference. Has anybody else seen anything similar? Or tell me that I'm barking up the wrong tree? Constructive criticism welcomed! Blue skies... Todd ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ Clamav-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/clamav-users
