I have seen this exact scenario crop up a few times recently with one of our servers. Also like you, there is nothing indicated in the Windows event logs or the FM server logs. I don't have any advice off hand though.
What is your setup in more detail? Mine is FM Server Advanced 11v5, Win2003 running on a VMWare server. We do not have any Custom Web publishing dB's, but do have a couple published via IWP. We have other servers that have web publishing turned on, but don't have any dB's being accessed via IWP or CWP. No issues with those servers.
This same server has been publishing the IWP dB's for quite some time (at least a year) with no errors. We then started gettting this situation come up about 8 weeks ago. We shutdown FM server cleanly and then reboot the entire Win2003 server when it happens. Note at this time, it was running 11v3 of FM server. Once the first occurence happened, it started happening more and more often until we had it happen twice in one day.
At that point, we decided to fully uninstall FM server and then re-install using the latest 11v5 version. This seemed to work in that no errors were encountered for about 2 weeks, but then it hit again. We did the reboot and are monitoring things. So far, we are at another 2 weeks without error.
In my case, there is an IT department that periodically applies Windows updates to the servers. My guess is that one of these patches changed something that the FM web publishing engine is not happy with. That was one of my main reasons for updating to the 11v5 release to see if it would help. Had been planning to do this anyway.
If we get this error again in the near future, I'll start digging in more detail as to what patches were applied to the server starting from the time we got the first error.
Thanks for chiming in Doug!
It sounds like our environments are very similar. I'm running one CWP page and it sounds like you're running a handful of IWP pages but otherwise very much alike.
What kind of anti-virus are you running? We are using ESET and excluding all FP7 files from the scan but I've wondered whether some of the pattern updates that come out about every day are causing the problems. I believe we tried uninstalling ESET a couple months ago and still had the problems so I don't think virus scanning is the cause.
I'm hoping to move to FM12 soon on a new server with a newer OS which may solve the problem so if this problem doesn't get resolved I'm not sweating it too much.
We made experience that CWPE outbursts happened when search spiders re-crawled the page with a wrong URL (query parameters incomplete or false). Check your Web access logs if that might be the case.
Von meinem iPhone gesendet
Doug's statement about IWP may give you a clue. We had our web vulnerability team scan our network some time back, and one of the things they found was similar to what Martin pointed out - that incomplete parameters submitted to the IWP page tended to cause a timeout. If the server is trying to chew through a bad parameter, it may end up grabbing up all the resources, and we have seen instances where we had to force-quit the WPE and restart it before.
There might be a connection.
Thanks for the feedback. We are running a McAfee product for anti-virus. We have provided info to the server team about what to not scan and was assured that it is setup accordingly when the server was deployed. Much time has gone by since then though. Maybe new staff on that team have changed parameters without knowing any better, etc. Given the comments below from Martin and Mike, I am going to ask them to double-check this.
I have no idea if we have any processes crawling our web pages. This environment is with a pretty large banking organization, and the network is pretty much locked down from outside world. We have IWP accessible dB's open on other servers, so it would seem odd that just the one server is experiencing this.
I am leaning towards anti-virus scanning being a possible issue.
Thanks again to everyone for the ideas.