We are working on a project that will require 100+ file triggers and the setup uses a VIP (F5 load balancer).
We recommend that:
- Install PTF RO94300:
If an agent cluster is used for File_Trigger jobs, the agent name used
in the message sent to delete a file trigger entry to the agent contains
the agent cluster name instead of the name of the agent actually
monitoring the files. Accordingly, when a File_Trigger job is being
completed, a delete file trigger message may be sent to a wrong agent
causing file trigger entries to never be deleted. This causes agent
initiators on the right agent not being released and eventually stops
the agent from accepting new requests. This fix corrects this problem.
- Increase RETRYINTERVAL (like the whole cycle of the round robin) for related agents in AGENTDEF table, so that ESP will not send duplicate requests.