hi there,
I know this has been analyzed in various blogs and forums, but I still haven't found any valid solution for this. Let me get more specific: Very often (at least 9-10 times day) we are getting workflows terminated, due to the following error in the logs:
OWSTIMER.EXE (0x14E8): Error in persisting workflow: System.Transactions.TransactionAbortedException:The transaction has aborted. ---> System.TimeoutException: Transaction Timeout
We have increased event throttling and transaction timeout in our web application, and the error doesn't occur when the process that is running the workflow is IIS. However, after a delay activity,when the workflow is picked up by the timer service of our application server(you will notice that the source process of the error is ALWAYS the timer job/owstimer.exe) , our workflows very frequently crash with the above error. I think it's not coincidence that the error happens aboutone minute after it gets picked upby the timer job, which probably means that we should increase the transaction timeouts for the timer process also (the default is one minute so 99% this is the source of our problem).
As I wrote above, I have already increased the <system.transactions>entry in the web config of our web app, but this setting seems to be ignored by the timer process of our app server. I also tried to add it in the owstimer.exe.config file to no avail. So the real question is:How can the system.transaction timeout value be changed so that it affects the workflow timer job of the sharepoint timer service (owstimer.exe)?
Any help would be highly appreciated because this has turned to be a really annoying problem for our users in the production farm - imagine having to restart ~10 workflows during the work day.
Thanks,
Spyros
