Re: [SQU] Logfile Rotation.

Palmer J.D.F. wrote:

> Hello,
>
> I am running 2 caches, My main cache is running squid 2.2 stable 5 on
> RedHat 6.0 2.2.17 kernel.
>
> I have just rebuilt my backup cache which is now running squid 2.2 stable 5
> with RieserFS and Async-IO (16 threads) on RedHat 6.2 with 2.2.17 kernel.
>
> I have an entry on both machines in the rc.local file which runs squid on
> boot up. The problem is that on the rebuilt machine the logfile rotation
> command (squid -k rotate) does not work unless I kill the squid processes
> and restart squid.
> By this I mean I kill the instance of squid that is started up at boot time.
> Once I have restarted it manually it works fine, and I don't have to kill it
> again.
> This command works on the older machine, from boot up.
>
>>From the squid.conf file I see that async-IO requires squid to send
> different term signals to the process in order to rotate the files, could
> async-IO be part of my problem?
>
> Can anyone shed any light onto this for me, I'd be very grateful.

If you are using a Red Hat style RPM, then you will have to manually (or
via the systems logrotate tools) move the old log..then run rotate.
This is normally automatic on Red Hat systems (rotations come once a
week by default...which can be altered in your /etc/logrotate.conf file.
  But if you want to force it, that's how you would do it.

If you want to go back to the original Squid way, just edit your
squid.conf logfile_rotate directive to be some number other than 0. The
official squid default is 10.

Hope this helps. BTW- WRT to the signal issue in async Squid, the -k
rotate command is not effected...only using the numbered signal is (i.e.
kill -USR1 <squid-pid>). So use the squid -k signals, and don't worry.
                                   --
                      Joe Cooper <joe@swelltech.com>
                  Affordable Web Caching Proxy Appliances
                         http://www.swelltech.com

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html