We installed dermio in a k8s cluster and the k8s cluster is planified to be down beween 1h30 am and 7h30 am.
This morning at 6h42, the planified refreshing reflection task is failed with the following error:
UnsupportedOperationException: Unable to use pdfs when no executors are running.
Query failed as Dremio was restarted.
Our dremio cluser is composed by 3 nodes: 1 coordinator and 2 executors.
At 8h33, the down time is finished and I got another message error:
The Default engine is not online.
At the same time, I found out one of the executor is shutdown automatically.
Our configuration is as following:
coordinator : heap size = 8Gb, direct size = 8Gb
executor: heap size = 8Gb, direct size = 8Gb
We are using Community Edition actually.
Why a reflection refreshing task can be launched during the server down time?
Do you have any idea for these errors?