Reasonable expectations of sysadmins

[Garance A Drosehn]

On 10/11/11 3:52 AM, Dag-Erling Smørgrav wrote:
> Mike Brown<mike at skew.org>  writes:
>    
>> Also, sometimes things go haywire after a reboot, especially after extended
>> uptime and updates to the kernel or core libraries, so I'm in the habit of
>> only shutting down when necessary. So if I don't see "and then reboot" in an
>> update procedure - and most of the time, security updates don't require it -
>> then I don't do it.
>>      
> Actually, this is an argument in favor of rebooting regularly, or at
> least after every major change, so you know the server will boot
> unassisted if something happens (power outage, cleaning staff tripped
> over the mains cable, etc.)  I once spent an entire evening coaxing a
> mission-critical database server back up after a simple disk replacement
> because a predecessor had performed an in-place system upgrade without
> verifying that the new configuration would boot cleanly.
>
> DES
>    
FWIW: If I have a production server which has been up and running for
more than six months, I often reboot the machine *before* making some
significant change, just to make sure the machine is still in working
order before I make that change.  I then make the change, and reboot
again.  There are times where I have discovered problems in that first
reboot.

(also note that in my case, most production servers which have been up
for more than six months have probably been up for more than a year)

-- 
Garance Alistair Drosehn            =   gad at gilead.netel.rpi.edu
Senior Systems Programmer           or  gad at freebsd.org
Rensselaer Polytechnic Institute    or  drosih at rpi.edu