We face this issue on and off with our IdentityMinder application which is running on a cluster of two servers. We frequently run into issues with tasks getting stuck in a 'In Progress' state.
I have had a look at the tech articles http://www.ca.com/us/support/ca-support-online/product-content/knowledgebase-articles/tec498092.aspx & http://www.ca.com/us/support/ca-support-online/product-content/knowledgebase-articles/tec615197.aspx. While they speak about how to provide access for tasks to be resubmitted individually, I need to know a way to resubmit a mass number of tasks. To illustrate with an example, on some days we have thousands of tasks getting stuck it is not possible for us to resubmit 3000-5000 tasks individually.
I need to figure out a way to -
a) Resubmit tasks in a 'mass' manner
b) Find the root cause for these tasks getting stuck and provide a long term fix
On running the following query on our DB server - select * from tasksession12_5 WHERE (state NOT IN (128, 512)) I get almost 450K rows. I don't necessarily need to resubmit all these tasks, I need find a way to resubmit tasks stuck for a selected date.
Also we have the following row counts for the various tables in the Task DB. I understand that these are really high in number, but is there a specific rule based on which I need to calculate what is the max limit on the number rows for the below tables before I purge these -
tasksession12_5 - 722715 rows
objects12_5 - 5519587 rows
lock12_5 - 5351804 rows
event12_5 - 4076086 rows
Note : This is from one of the DB servers, the other DB server has more or less a similar number.