How to Upgrade an Existing Cluster
To upgrade an existing cluster while maintaing availability, perform the following steps for each node of the cluster:
1. In your current environment:
a. Check the cluster health using the following command:
adaopr dbid=dbid display=cluster
All nodes must be in status SYNCED.
b. Shut down the node. Always shut down the primary node last.
c. Back up the database (optional).
2. In your new environment:
a. Convert the node with the following command:
adacvt dbid=dbid convert=vNN
b. Restart the node to rejoin it to cluster with the following command:
adastart dbid
All changes in the cluster are recorded as writeset and transferred to the restarting node via 'Incremental State Transfer (IST)'.
Note:
An attempt to connect a node with an 'old version' to a group already using the 'new version' could be rejected with the following message:
ADANUC-E-WSREP... Group requested repl_proto_ver: 11, max supported by this node: 10.
Upgrade the node before joining this group. Need to abort.
This works as designed, you can only connect a node with a 'new version' to a group still using the 'old version', but not vice versa.