Details
Description
1. Create 3 node cluster
2. Bring 1 node down
3. Using CLI graceful failover the down node
In Step 3, CLI allows graceful over and rebalance fails. In UI we do now allow this, only hard failover for a node that is down. The CLI should have similar behavior as the UI
Example: In the example, we have a 3 node cluster 10.6.2.144, 10.6.2.145, 10.6.2.150 (we stop this node in process) and then use CLI to do graceful failover. See how it happens and rebalance fails
[root@palm-10308 bin]# ./couchbase-cli failover -c 10.6.2.144:8091 --server-failover:10.6.2.150:8091
ERROR: option --server-failover:10.6.2.150:8091 not recognized
[root@palm-10308 bin]# ./couchbase-cli failover -c 10.6.2.144:8091 --server-failover=10.6.2.150:8091
ERROR: Username cannot be empty. Please use -u or set environment variable CB_REST_USERNAME.
[root@palm-10308 bin]# ./couchbase-cli failover -c 10.6.2.144:8091 --server-failover=10.6.2.150:8091 -u Administrator -p password
INFO: graceful failover
Rebalance failed. See logs for detailed reason. You can try rebalance again.