Github user IgorBerman commented on the issue:
https://github.com/apache/spark/pull/20640
@squito In general I agree with you, however current status of hardcoded
max 2 failures is blocking, so I've created simple fix that will demand less
integration testing. I can close it.
I think in context of MesosCoarseGrainedSchedulerBackend the
TaskSchedulerImpl called scheduler so to test if slaveId is blacklisted
probably following is needed:
{code}
!scheduler.nodeBlacklist().contains(slaveId)
{code}
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
