dan goldberg -
January 19, 2009
The batch database processing errors related to the following scenarios which were causing user processes crash to have been resolved:
1) "Transaction (Process ID ###) was deadlocked on lock | communication buffer resources with another process and has been chosen as the deadlock victim."
2) "Thread was being aborted."
The first error was due to database table locking. This popped up because we are starting to get a lot of usage and this revealed that our transaction tracking tables were being read/written in inconsistent orders.
The second error was due to incorrect handling of application recycling caused by IIS. Trapping for and handling this error, we are now able to resume processes once IIS recycles the application pool.
We apologize for any inconveniences these errors may have caused. If you run into either of these errors in the future (or any others) please let us know so we can get right on them.