[EMAIL PROTECTED] escreveu:
I'm not certain, but I think the problem is that you're creating the log
file while freshclam is still using it. The "create" command in logrotate
says:
Immediately *after* rotation (*before* the postrotate script is run)
the log file is created (with the same name as the log file just
rotated).
Sending the HUP signal to freshclam after the create isn't what you want.
Freshclam just closes its connection to the logfile upon receiving the HUP
signal, and opens a new one upon first write.
I think that's correct ...... this is, by far, the most used log
rotation procedure.
logrotate 'rotates' logfiles: rename all the old files and increase
their numbers, remove the oldest one and rename the actual freshclam.log
to freshclam.log.1. Logrotate, then, creates a new freshclam.log with
permissions and owner/group configured. Until now, freshclam doesnt know
what's happening. File descriptor will continue to point freshclam.log.1
file and not our new freshclam.log. If nothing happens, freshclam daemon
will continue to write on it's filedescriptor, thus writing information
on the 'old' logfile and not the new one. So, there goes the HUP .....
it will instruct freshclam daemon to close and reopen logfiles, thus
getting the filedescriptor on the newly created freshclam.log,
I dont think that's the problem, at least this is the procedure i've
seen a LOT of daemons using for log rotations. And i took these
freshclam logrotate file from the official RPMs, i have to suppose they
are correct :)
Anyway, thanks for your answer !
--
Atenciosamente / Sincerily,
Leonardo Rodrigues
Solutti Tecnologia
http://www.solutti.com.br
Minha armadilha de SPAM, NÃO mandem email
[EMAIL PROTECTED]
My SPAMTRAP, do not email it
_______________________________________________
http://lurker.clamav.net/list/clamav-users.html
