[Samba] Re: Best practices for long-running Samba server

[David Landgren]

On Thu, 27 Jan 2005 14:14:46 -0800, Spike Burkhardt
<burkhardt.richard at ssd.loral.com> wrote:
> Paul & David,
> 
>  Good points that I agree with but at least at three managers I've had want to the uptime get bigger & bigger.  Something about the 99.95% uptime
> "industry standard." :-(((  Ridiculous.

They are wrong.

Tell them gently that they should not confuse uptime with
availability. A planned reboot at 03:00 that has no effect on the
availability of the server is much better than an unexpected outage at
14:25 because a leaking database backend ate all your shared memory
segments. That does wonders for availability.

If you can't trust your machines to reboot themselves unattended and
bring themselves back up to operational status then you have shall
have to bring a pair of pyjamas into work and set up a cot in the
machine room.

99.95% is about 4 and half hours per year. Ask your managers to cover
you for the allowance of having a server being offline one whole
afternoon per year. Ask for it in writing.

David