We have an Oracle RAC cluster with 8 threads each for two nodes. It appears that the Oracle driver (non-spatial) runs out of threads on the first node, but never falls over to the second node.
Is there a way to configure the FME workflow to do this, or perhaps the driver itself? We have intermittent failures during our overnight process that magically "fix" themselves the next time this workflow runs. And after a lot of testing and trial-and-error, we came to the conclusion that the 8 threads on the first node are probably filled by other processes, and FME gets denied access because it never goes over to the second node. The error we get (not exact wording) is that the data source is unavailable.
Has anyone experiences this and has a solution or a hint at what might help?
Thanks!
