Re: [squid-users] Did logrotation code change between 3.1.9 and 3.1.12?

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