WSS/MOSS 2007 causes IIS some heavy duty pain. Put it up, leave it up, and eventually, it will die. We've all seen it, and I'll bet those of you who are lucky enough to run SP1 or other patched up systems have seen it too. Hopefully we will get all patched up soon too, but I don't expect that to make things perfect.
The unexpected crashes of IIS seemed to do the most damage to the indexing process. The indexer itself can take a hit and stop functioning (you probably will see a lot of junk in the event logs) or the web front ends it is trying to use might unexpectedly crash. We got slightly toasted (not quite burned) by this several times.
Since we started rebooting the farm Saturday night (before Sunday's full crawl), things have been better (I would hate to say perfect, but it has been perfect).
I know that server folks don't like prophylactic reboots. They really would rather address the problem, but since some of what ails SharePoint servers are real bugs, memory leaks, etc., this kind of thinking has to give way to good, old fashioned, weekly reboots.
For those that do like to live on the "wild side" and patch their servers, check out FIX: You may be unable to manage IIS by using Server Manager if two threads access IIS at the same time as an alternative to having IIS repeatedly crash.
ReplyDelete