Re: OT: Windows Server 2003 uptime

["Robert G. (Doc) Savage" <dsavage@peaknet.net>]

On Tue, 2004-01-27 at 18:12, Steven Pritchard wrote:
> On Tue, Jan 27, 2004 at 05:29:18PM -0600, Ray Lewis wrote:
> > Btw, if you do the math that's not bad for a windows server. <grin> 
> [...]
> > The system uptime is 5761156 seconds.
> 
> So almost 67 days?  Wow, they're finally getting up there with what my
> box would do with a 1.0 kernel.  ;-)

Great catch, Ray. 5.7 megaseconds. Big hairy deal :-) 

This may be a bit subtle of a point for some, but consider what these
uptimes represent as fractions of Windows' and Linux's supported
lifetimes. There are many documented examples of Linux systems running
nonstop for periods exceeding 20% of their supported lifetimes. When you
think about it, this is a truly extraordinary number.

I have a client in Belleville whose DNS server running RHEL/ES v2.1 hums
flawlessly for six months or so between kernel errata upgrades -- the
only times it is ever taken down for two minute reboots. That's better
than 99.999% availability, without clustering!

By contrast, thirty days is about as long as I normally allow any of my
Windows 2000 (or Sun) servers I administer to run before rebooting them
to recover the heap and swap space that inevitably leaks away. There may
be examples of Linux boxen with such slow leaks, but I personally
haven't seen any.

--Doc


-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.