Be forewarned: I just installed the latest udev update onto two of my CentOS 4 servers and it caused a kernel panic. After resetting the servers, they have come up fine. I am still researching as to what caused the problem.
UPDATE: I believe the problem was caused by an outdated kernel. We were running 2.6.9.
Today I ran yum update on my mail server and noticed that the ClamAV packages were set to be upgraded. Trusting rpmforge’s packages as I do, I ran the updates and attempted to restart clamd and clamav-milter services. This is where my problems began. When attempted to start clamd I received the error:
ERROR: LogFile requires full path.
When I reviewed the clamd.conf file the LogFile line read:
LogFile unix:/var/log/clamav/clamd.log
I’ve never seen the unix: prefix in a path before so I removed it. When I did clamd started right up. Then I looked that messages logfile and saw the following line:
clamav-milter[19418]: No clamd server appears to be available
I looked in the clamd.log file and saw that is was set to LocalSocket /tmp/clamd.socket, but the clamav-milter.conf was looking for clamd’s socket at unix:/var/run/clamav/clamd.sock. So I changed the ClamdSocket line to unix:/tmp/clamd.socket (to match clamd.conf) and restarted clamav-milter. Then the error messages went away.
Apparently the 0.95.1 version also turns off the X-Virus-Scanned headers as well. I turned that back on in the clamav-milter.conf file.
Apparently there was a major architectural change in ClamAV 0.95.
UPDATE: A new set of packages was released this week. I don’t know if they fix the problem, but I suspect they caught their mistake.
Posts
All Comments
