[PKG-Openstack-devel] question about cinder logrotate scripts

[Thomas Goirand]

On 08/27/2014 10:49 AM, stanzgy wrote:
> Hi all, I have noticed that our cinder service will get restarted every
> logrotate runs.
> IMO it's very dangerous to do this kind of things.
> Is there any specific reason to do this? My suggestion is to remove the
> postrotate section in these scripts.
> 
> /var/log/cinder/cinder-api.log {
>     daily
>     missingok
>     compress
>     delaycompress
>     notifempty
>     postrotate
>         if dpkg-vendor --derives-from ubuntu ; then
>             service cinder-api restart >/dev/null 2>&1 || true
>         else
>             invoke-rc.d cinder-api restart >/dev/null 2>&1 || true
>         fi
>     endscript
> }

Hi,

Sorry for the delay replying to this.

I'm not sure which version of Cinder you're using, but since 2014.1-6,
the logrotate scripts are using copytruncate and do not restart daemons.

Cheers,

Thomas Goirand (zigo)