On a private mailing list Richard Hundhausen dug up this good info if you've ever experienced the "Deadlock detected" error in ASP.NET, perhaps when calling a long running Web Service:
This reply from David Wang (MSFT) on the Google groups with some interesting information:
You need to turn pinging back on. Report unhealthy depends on failing the ping (and failing the ping is what triggers IIS to recycle the worker process). Enabling pinging should not affect your other recycling options, since all you needed to do was turn off all the time-based recycling metrics -- from a default installation, all you need to turn off are "Idle Timeout" (15 minutes by default) and "Periodic Recycling" (29 hours by default).Regarding - responseRestartDeadlockInterval and responseDeadlockInterval -- there isn't anything comparable in IIS6 because ASP.Net is using the "report unhealthy" mechanism to get IIS6 to recycle itself. How the interval is configured should be specific to ASP.Net.
You need to turn pinging back on. Report unhealthy depends on failing the ping (and failing the ping is what triggers IIS to recycle the worker process). Enabling pinging should not affect your other recycling options, since all you needed to do was turn off all the time-based recycling metrics -- from a default installation, all you need to turn off are "Idle Timeout" (15 minutes by default) and "Periodic Recycling" (29 hours by default).
Regarding - responseRestartDeadlockInterval and responseDeadlockInterval -- there isn't anything comparable in IIS6 because ASP.Net is using the "report unhealthy" mechanism to get IIS6 to recycle itself. How the interval is configured should be specific to ASP.Net.
Ads by The Lounge