Log rotation - Graceful restarts....

I was hunting through additional settings in webmin and just noticed under my logfile rotation.

I have the following entries for commands to be run on a weekly basis.

(TLD of another user) Set on a new account created via Virtual Min - weekly
/usr/sbin/apachectl graceful

(Sub-domain of mine) Set on an account created via cPanel Backup import - weekly
/usr/sbin/apachectl graceful

(Sub-domain of mine) Set on an account created via VirtualMin as a Sub-domain of my TLD - weekly
/etc/rc.d/init.d/httpd graceful

These are the ONLY 3 graceful restarts I can see…
Should I remove these? I don’t want to destroy the system by sending it 3 graceful restart requests on a weekly basis to start with…


The “graceful” restart there is used to close out the old logfiles, and start writes to the new ones.

A graceful restart is easy on Apache – no open connections are aborted, and a config test is performed to check that the configuration file syntax is okay.

So this would generally be considered a pretty safe way of handling log file rotations.

I haven’t heard of anyone having problems during the logfile rotations… is there a problem in particular that you’re concerned about?


I just like to see the service still running as in my mind that means it’s going A’OK.
Just monit sees the graceful restart and the “uptime” is reset >.<

Thing is I have 3 seperate graceful restarts. I think I might edit them to all be under the one so it only happens once a week.

On that note though. Wouldn’t it be better if Virtualmin Appended log rotations for servers to the one? So that it does it all at once and not several times depending on when the rotation for that server’s logs kicks in?

So back to this.

Does virtualmin create a new entry in log rotations for each individual server?
Just I can see if I have 30 hosts … that’d be 30 graceful restarts a week…

Would it not be more logical to have a single log rotation process that calls on an external file that virtualmin will edit/update for commands to run as servers are added/removed?

I feel that would be much more efficient.



Should Virtualmin be creating new individual log rotations per new virtual server or should they all just be done at once.

As I have 3 graceful restarts in there at the moment… I would rather have 1 global restart happening.


Well, it’s working the way it’s been designed to… if you have concerns over the reason it was setup that way, or have recommendations of another way to do it, I might suggest filing a support request using the Support link above.