In a customer, some months ago I was facing a problem related to Windows 2003 SP2 (SNP enabled), old Broadcom drivers and Exchange HTTP clustered resource failing. It was resolved upgrading drivers & disabling TCP Chimmey.
After this, no problems in 2 months but now 2 failovers with the HTTP resource involved in. I´ve seen no events explaining this error (only the obvious "...the resource HTTP 100 failed").
Customer is using RPC over HTTP (Exchange backend is the RPC proxy) for 800 clients. Looking at the appication pools, I´ve seen the RPCoHTTP is configured to the DefaultAppPool, and the customer changed the default recycle time to 6:30 am everyday. In this M$ document http://technet.microsoft.com/en-us/library/cc540453(printer).aspx#RefOutDir
, I´ve seen that is recommended to move RPC over HTTP to a new application pool without recycle.
Daily, we can see an event like this:
"A process serving application pool 'DefaultAppPool' exceeded time limits during shut down"
Any ideas to cope with this error??