Our standing policy was Quarterly unless a patch required it; which we do monthly. Lately it's been every month that we reboot though...even though no one (read: staff) seems to have a good reason why; management now seems to expect it..
Just a warning, but I patched my servers last week. On reboot one of them didnt! It claimed msgina.dll was corrupted & simply would not boot the OS until I restored the file. After getting it to boot it had corrupted asp.net & the input locale, which we had to reghack to repair. Even now there are strange entries appearing the event log..
Ironically I dont even know which patch it was as we roll them out with WSUS! I have 3 other servers waiting to be rebooted, but I fear the worst if I do!!
However, some application may not release resources (PTEs, nonpaged or paged pool, working set or private set memory, etc...,etc...)This is known as leaking....
Tivoli and Domino server are two apps I know of that will constantly leak working or private set memory. When this memory usage gets too high, your server will slow and may dump. (Tivoli's documentation used to actually say that this is normal. I don't know if that is still the case)
I encourage admins to use perfmon every so often to see if any apps are leaking. Then you can be on top of problems that may occur.
Still, rebooting isn't the solution; it's a workaround. The solution is to contact the vendor of the leaky application and have them fix the problem they created. If they can't or won't, find another vendor that will.
You are absolutely correct xmsre. If you have a leaking app, the best thing to do is to get a dump of that app and have the vendor or MS look at it. When you get a dump with a perfmon log, you can track down the thread stack that is causing the leak...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.