Jon Stanley wrote:
I'll throw a -1 in here (not that my vote matters for anything :)
).
Not because of any risk in the change, but rather it violates the
concept of a change freeze. It doesn't affect service, so if I went to
our change management meeting at work with something like this during
a change freeze I'd be laughed out of the room and told to wait.
When such a thing happens, and it does, I wait for the big outage I
can't recover from without changing stuff during a change freeze, and
drag their sorry asses back in the conference room.
First thing on the agenda of course is appending "> /dev/null" to the
condrestart of some logrotate file, not the outage.
-Jeroen