edi tracking table

I’m looking for information on strange issues noticed after an out of memory error and a subsequent heap dump. Usually these occur if the file system runs out of allotted space. (AIX)
IS and TN 6.1
We have seen tasks in TN which remain in pending state (not all, just some)
Some scheduled services in IS do not run, these are very short running services such as scheduling an FA report or a document count. (may be covered in IS_6-1_Fix50)
more recently, when running the fa report, the error "com.wm.app.tn.err.EXMLException: Error getting EDI_tracking data " was returned.

re-starting the repo server in each case has resolved the issue, but this is not an ideal solution.

I havent seen any postings here or on advantage. I’m just looking for anyone with similar experiences or resolutions attempted.

the issue with reading the error tracking table is a mistake. this should not have been included in the issues.

the real issue, why would the repo server need to be re-cycled after a memory dump? how can we recover without stopping the IS on all nodes of the cluster, and then re-start the repo server. What is broken when the heap dump happens?