Some incidents that we face are expected. Usually we wait for problems when something changes in an environment.
But, some times, for no apparent reason, with no systemic alteration, we encounter errors where our first reaction is: what a f ***!?
This time we find a java exception in a standard domain for GoldenGate Director;
For months the application behaved stable and functional, until it did fails for no apparent reason;
When I saw part of the exception, “XAER_RMFAIL: Resource manager is unavailable” went straight to talk to one of the best DBAs know that Matheus Boesing, to request a check on the resource manager database, (No problem found).
… then we fall back into a BUG: 11672297 Bug: ORA-01092 MAPPED TO XAER_RMERR instead of XAER_RMFAIL – (Doc ID 1329800.1)
In version 12.1, this bug is fixed, but as a palliative solution can do the following:
Increase the value (Maximum Duration of XA Calls) in JTA configurations of weblogic domain, the default value is 12000, in my case, I adjusted to 48000;
The problem was solved, at least for now.