Lately in our production environment we saw that when some of our services where called in rapid succession, some of them failed. Under normal load we never had issues but under bigger load we saw the following error in our BPEL instances:
Failure in Oracle WSM Agent, category= security, function=agent.function.client, stage=fault due to RuntimeException.
java.lang.IllegalArgumentException: Context of the same RID, 0:1:4 already present in this family, f42da30f-5c89-49d9-a4fd-bd2b7e80644e-00154a3f.
It also seemed that those faults where creating stuck threads and after a while making our managed servers unstable and unreachable. After some digging around we found that there was a patch for it. See Sporadic error during invocation of a SOA composite, ‘Context of the same RID, xxxxxx already present in this family’ (Doc ID 2150992.1). It was released since june 2016 but it seems it was never put into a bundle patch somewhere. After applying the patch the errors where replaced by warnings:
Sep 13, 2017 11:48:22 AM CEST Warning oracle.dms.context BEA-000000 Duplicate RID detected. Multiple tasks are attempting to execute with the same exec-ctx identifier 9f31c77a-4bee-4f57-a970-9275a6577bcb-00030ae5,0:1:4. Will use the unrooted identifier 9f31c77a-4bee-4f57-a970-9275a6577bcb-00030ae5,1:32171 for this task.
and we had no more stuck threads and all our instances where running fine again.