Hello,

Would you be able to provide the Job logs for the two successful runs?

Is this the only workspace affected?

Is this on a schedule or automation?

Are you triggering the job through the FMEREST API? Could it be that the front end is expecting a call within 5 minutes and if it doesn't get a response, try it again?

This is the only workspace affected and this workspace is just ran through the workspace app link (.../fmeserver/apps/name). It is always starting the duplicate exactly after 5 minutes. I inspected the network requests and it is pending all the way until the duplicate is finished.

Unfortunately I am not able to share the job logs.

But here is explained what is happening during one minute before the duplicate starts:

• Reading data from different tables of our database with FeatureReader
• 64 rows marked in the log with the same exact second (52 seconds before dup), most of them marked as "STATS"
  • coordinatesystemsetter, Vertex creator, reprojector, coordinaterounder
  • Testers
  • ListBuilder
  • AttributeManager
  • AttributeFilter
  • Sorter
  • Excel Reader: Closing dataset?
  • FME API version of module 'ArcSDE' matches current internal version
  • Performing query against ArcSDE dataset
  • Connection made to server X for dataset Y
• Then for the rest of the minute FeatureReader is reading table/feature class and continues reading after the duplicate is created.
• (And ofc there is not mention anywhere indicating that new workspace would be started)

So far I've compared line by line the logs when the workspace is ran locally vs the workspace ran from fme server and there and there seems to not be anything out of ordinary.

I've also tried to increase the time for some fmeServerConfig.txt settings, with no results (it still starts a duplicate exactly at 5 minutes):

RELAYER_MAX_PUBLISHER_ACCEPT_ATTEMPTS=30000 -> 33000

CONNECTIONPOOL_EXPIRY=300 -> 330

SOCKET_HANDSHAKE_TIMEOUT=30000 -> 33000