After the upgrade to 11.5.2 the AP Log Window did not display any messages for over 6 minutes. AP was up and appeared operational but no messages were displayed. After six minutes the AP log switched and message were displayed normally. Investigating, we found an abend trace messages in the logs and a dump file. The dump file report appears to indicate a thread deadlock condition may have caused the abend.
Examining the diagnostic report, we that the failure is not caused by a deadlock condition as indicated by the dump file. Although there is one thread initiating the shutdown and dump after waiting eight minutes for a resource already held by another thread (which is normally triggered by a thread deadlock condition), that thread is not waiting for another resource, which would be held by the previous one or some other thread in case of a deadlock. The thread blocking the resource is instead waiting for a database table to be deleted. We think, that it takes so long, because some NSM processes are taking most of the computational power on that machine at a given time after restart. This is unfortunate althou not entirely unnecessary. There is a paragraph in the Install Guide section labeled as Server Requirements describing that while small volume Automation Point could run on single 2GHz processor, a high load system or system running NSM has to have two 2.4GHz processors. You happen to have one 2.4GHz processor running on this system.
Dedicate one more 2.4 GHz processor core to the machine (VM) to meet the minimum requirement for a high-load system.