FDMEE Batch Scheduler--Why Did This Run Twice?
If you’re familiar with EPM and FDMEE, you know that you can schedule batches on load jobs, to add some automation into your data load process. This has been a feature all the back through FDM Classic, and continues into the latest version in EPM 11.2.x.
Recently, we had a client who uses a number of batch jobs encounter an issue where they would see each job run twice. This only occurred in their Production environment (where we have two FDMEE nodes configured) and not in their Development environment (where there is only one FDMEE node). After working with Oracle, they recommended we look into Doc ID: Document 2735938.1.
This document has the following verbiage:
A system administrator needs to follow the below steps to modify the server configuration:
Change the socket provider from UDP to TCP.
Create a simple TCP/IP echo service to enable port 7. 3. Enable ICMP inbound/outbound
Well, that’s clear as day, right? Given that this was a PAAS environment hosted in the Cloud, it was very possible that a firewall rule could be preventing communication on Port 7. However, after reviewing with the architects, it was determined that wasn’t the case. Instead, we found that this ‘can’ be a function of Windows Server 2019.
Microsoft is slowly removing some of the older backed-in TCP utilities and making them optional. In this case, we needed to enable Simple TCP on the servers which allows for ‘echo’ calls over to Port 7. In order to do this, simply do the following:
Go to the Start menu and look for Control Panel to open it.
Navigate to Programs.
Select turn Windows feature on or off.
Select Simple TCPIP Services.
Apply changes.
Close.
Once the above is complete you will need to stop EPM services and restart the servers, and your duplicate FDMEE jobs should go away.