We were experiencing job reload issue, and we discovered it might be due to our scheduler node not being passed the reload tasks from the central node. Originally, we had our setup such that the central node had a scheduler on it and was set to manager and worker with a max of 4 concurrent reload tasks, and our scheduler node had a scheduler on it and was set to worker with a max of 4 concurrent reload tasks. The QMC said all appropriate services were running on the scheduler node, but if more than 4 reload tasks were trying to run simultaneously, we noticed the first 4 would be handled by the central node, and the 5th one would be queued. That 5threload task was never reaching the scheduler node.
When we tried to set the central node to manager only, no reload tasks would run at all, and the QMC would give error no slave scheduler found.
