Re: How Frequent to Reboot Server Best Practices?
- From: Jason Gurtz <jason@xxxxxxxxxxxxxxxx>
- Date: Tue, 16 Aug 2005 11:42:13 -0400
On 8/11/2005 13:54, Joel wrote:
> I'm wondering if there is any documented best practices for how frequently
> to reboot a Windows Server 2003?
If a server has to be rebooted on a scheduled basis for other than patches
then you have a situation where there is bad hardware or bad software on
the machine. Typically, the bad software would be a driver or some poorly
written 3rd party vertical market application that's not part of the OS...
*ahem* often requiring the server to be running *LOGGED IN* at all times! :/
A bare NT4 BDC file/print server can easily go 400+ days without
rebooting. I would expect that 2K and 2K3 machines are even more stable.
Often a badly behaved app that leaks memory or handles is a child of the
shell and can be band-aided by simply logging out and back in again. Use
of the Sysinternals utility Process Explorer can be very instructive. It
is more fun to badger the people that wrote the low-quality software until
they fix it.
Also, some services are badly behaved *cough* still IIS *cough* and can be
simply restarted every now and then. Even core services like WINS can be
problematic depending on load but I think a lot of that has been fixed
since NT4. Nonetheless, keep some baseline numbers around and script the
restarting of processes that grow too large over time.
There's no reason at all that you should feed UNIX people's laughing at
window's uptime. Win9x has been taking a dirt nap for a long time, send
cheers when Linux panics or BSD dumps core!
~Jason
--
.
- References:
- Prev by Date: Server Licensing Question
- Next by Date: Re: How Frequent to Reboot Server Best Practices?
- Previous by thread: Re: How Frequent to Reboot Server Best Practices?
- Next by thread: RE: Spooler unable to connect
- Index(es):
Relevant Pages
|