2

We have an IIS 7.5 web service on Windows Server 2008 that handles WCF requests from C# clients. This service is configured to have Maximum Worker Processes = 1, so it is not a web garden.

IIS is setup to recycle itself at the same time every day (3 AM). I am trying to debug gnarly connection issues, so I wanted to be sure the application pool was not recycling itself. I configured the pool to log an event when it recycles itself.

To my surprise, I see the following entries in the System event log:

Level: Information Date/Time: 3/23/2012 3:00:00 AM - Source: WAS - Event ID: 5076

A worker process with process id of '6636' serving application pool 'MyAppPool' has requested a recycle because it reached its scheduled recycle time.

Level: Information Date/Time: 3/23/2012 2:59:39 AM - Source: WAS - Event ID: 5076

A worker process with process id of '9364' serving application pool 'MyAppPool' has requested a recycle because it reached its scheduled recycle time.

IIS is correctly recycling the application pool at 3 AM. However, I do not understand why I would be getting two recycle events in the log within a few seconds of each other. The maximum number of processes is 1. Does IIS sometimes allocate multiple processes for an application pool that is specified as having 1 process?

-- edit --

I connected at about 4 PM today and only saw 1 w3wp.exe process. There are no other event log entries that would indicate a crash.

4
  • Are there any other events that indicate that a previous work process had crashed or anything like that? Have you looked at the process list, are there more than one w3wp process (or more than one per pool on your server)? Mar 23, 2012 at 20:56
  • Just 1 process, and I don't see any other suspicious event log entries. Mar 23, 2012 at 21:16
  • No, it doesn't. That looks like 2 consecutive worker processes quit because they thought they were reaching the time limit. Interesting and odd.
    – TristanK
    Mar 24, 2012 at 2:56
  • Exception being at recycle time - if overlapping recycles are enabled (default), 2 worker processes for the same pool overlap. Usually for single-digit seconds.
    – TristanK
    Mar 24, 2012 at 2:57

1 Answer 1

2

We have the same issue on several servers. If I reset the times they will only single recycle until the date changes, then it starts with the duplicates again. Microsoft told us this was because of time drift on the VM servers.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .