Details
-
Improvement
-
Resolution: Unresolved
-
Major
-
7.1
-
0
Description
In CBSE-13981 it was raised that the documentation isn't clear enough and that users need to be expressly informed that with the advent of Query's graceful shutdown applications could start to receive errors 1180 & 1181 during the shutdown process. The advice when receiving these errors is to retry such requests on a cluster node that isn't being removed, if possible, noting that the cluster manager will not report the nodes shutting down as being unavailable until the process is complete.
This constitutes a behaviour change from prior releases which applications should be aware of.
A 7.1 server can avoid reporting these errors by always failing over a node before removing it.
It will not apply by default in newer releases.