On 04/06/11 07:17, Rich Rauenzahn wrote:
My logrotating done by logrotated doesn't work anymore...
/var/log/squid/*.log {
weekly
rotate 52
size 100M
compress
notifempty
missingok
sharedscripts
postrotate
# Asks squid to reopen its logs. (logfile_rotate 0 is set in squid.conf)
# errors redirected to make it silent if squid is not running
/usr/sbin/squid -k rotate 2>/dev/null
# Wait a little to allow Squid to catch up before the logs is compressed
sleep 60
endscript
}
I end up with an access.log that has the wrong number and logrotate
can't compress it.
I think what is happening is logrotate renames access.log to
access.log.0, and then squid -k rotate renames it to access.log.1, and
then logrotate tries to compress logrotate.0, which it now can't find.
Or something like that. My uncompressed store.log is actually at .2
store.log i rarely useful. You may want to remove it from the config.
I've worked around it by setting logfile_rotate to 0, but I'm
wondering if this is a recent change with unexpected side effects when
working with logrotate.
That would be the problem. "logfile_rotate 0" is a requirement when
doing log rotation outside of Squid. Such as with logrotate.
It has been that way since Squid-2.6. The mystery is why it worked for
you before. Or why the squid.conf changed during the upgrade.
Amos
--
Please be using
Current Stable Squid 2.7.STABLE9 or 3.1.12
Beta testers wanted for 3.2.0.8 and 3.1.12.2
