Details
-
Bug
-
Resolution: Unresolved
-
Critical
-
None
-
Untriaged
-
Unknown
Description
Here are the steps for reproduction:
1. Create a cluster of 1 eventing node and 1 data node.
2. Deploy an eventing handler H1.
3. Rebalance out eventing node.
4. Delete source bucket of eventing handler.
5. Add an eventing node to the cluster.
6. Start rebalance.
Rebalance in step 6 fails. Eventing will try to spawn eventing handler H1 on newly added eventing node as it was never undeployed.
Since source bucket is missing, deployment will not go through and consequently
rebalance will fail. Eventing handler H1 will remain in bootstrapping state on the newly added eventing node and all the subsequent rebalance will fail.