Re: Systems administration guidelines
On Tue, Feb 22, 2005 at 01:58:09PM -0800, Alvin Oga wrote:
>
> On Tue, 22 Feb 2005, Fred Blaise wrote:
>
> > I am running several debian testing servers. I would like to find
> > guidelines/how-to/pointers to administer Debian servers the most
> > effectively possible.
> > ie: I wrote a script that run apt-get update && apt-get -y dist-upgrade
> > on a daily basis, sending me an email with the output..
>
> good
>
> > is it safe?
>
> depends
>
> > Should I do it or not?
>
> depends
>
> if it's not a critical system ... automating it is good ..
>
> if it's critical, you should test the patches before update/upgrade
>
> or gamble that somebody else can fix it while, you happen to be on
> vacation or sleeping
>
if it's critical enough that you have a second identical box waiting
for fast replacement/recovery, you might consider automatic update
_of_the_backup_box_ with automated regression testing after each
update. If the regression tests are successful, you can swap boxes
and the prior primary becomes the backup with the newly updated
being the new primary. But the box swap should be manual, so I'm
not sure what is gained.
Just a suggestion. I don't do critical servers, and I don't worry about
having the very latest tweeks on my systems.
--
Paul E Condon
pecondon@mesanetworks.net
Reply to: