On Thu, Jul 12, 2012 at 2:42 PM, Chiradeep Vittal <[email protected]> wrote: > According to > https://git-wip-us.apache.org/repos/asf?p=incubator-cloudstack.git;a=tree;f > =patches/systemvm/debian/config/etc/cron.daily;h=e28877633af747315ca525d866 > 208bc6f8bfae03;hb=HEAD > > > All log files in /var/log/cloud that are older than 7 days will be > deleted. > Logrotate has been configured to roll over other system logs (e.g., > HAProxy) >
So just a nitpicking tangent/question. Why do we maintain stateful data (logs) on VMs that are effectively designed to be stateless....should our preference become to send those off the VM to a syslog server (and make that a configuration option?) and not have any logs on the machine at all, or perhaps just a small subset (last 24 hours incase things go horribly wrong and network or syslog server dies) --David
