Adds a node back to the cluster after a failover
couchbase-cli server-readd [--cluster <url>] [--username <username>]
[--password <password>] [--server-add <servers>] [--group-name <name>]
[--server-username <username>] [--server-password <password>]
DEPRECATED: This command was deprecated in 5.0.0 and will be removed in future releases. Please use the man:couchbase-cli-recovery[1] subcommand which provides similar functionality to this command.
The server-readd subcommand is used to add a server back to the cluster. This operation is useful after a node is failed over and also when a node is removed from the cluster for maintenance.
When a node is failed over and removed from the cluster it may be able to be added back to the cluster. An example of this is when a node loses power. This node might get failed over and removed from the cluster, but once power is restored to the node you may want to add it back to the cluster.
Another use case is taking a node out of the cluster for maintenance. This is done by gracefully failing over a node to ensure there is no data loss. The administrator can then perform maintenance on the removed node and add it back with the server-readd command.
It is also possible to add a server back to the cluster without removing the data and instead having the server recover data from where it left off. This is called delta recovery and is available from the man:couchbase-cli-recovery[1] subcommand.
Note that after the server-readd subcommand is run you must rebalance the cluster. See the man:couchbase-cli-rebalance[1] command for more information on rebalancing a cluster.
- --server-add <servers>
-
A comma separated list of nodes to readd. The each server should correspond to the hostname or IP address of a server to be added back to the cluster.
- --server-username <username>
-
Adding a node back to the cluster can only be done by a user with the appropriate credentials. This flag specifies the username for a user who has the ability to modify the cluster topology of the node being added back.
- --server-password <password>
-
Adding a node back to the cluster can only be done by a user with the appropriate credentials. This flag specifies the password for a user who has the ability to modify the cluster topology of the node being added back.
- --group-name <name>
-
The name of the group to add the node to.
If there is a node at 192.168.1.6:8091 that you want to add back and that node has credentials that allow us to change the cluster topology with "Administrator" as the username and "password" as the password, run the following command.
$ couchbase-cli server-readd -c 192.168.1.5:8091 --username Administrator \ --password password --server-readd 192.168.1.6:8091 \ --server-username Administrator --server-password password
If you need to add back multiple nodes then you would run the following command.
$ couchbase-cli server-readd -c 192.168.1.5:8091 --username Administrator \ --password password --server-readd 192.168.1.6:8091,192.168.1.7:8091 \ --server-username Administrator --server-password password
If you wanted the nodes being added back to the cluster to be a part of the "East" group we would run the following command
$ couchbase-cli server-readd -c 192.168.1.5:8091 --username Administrator \ --password password --server-readd 192.168.1.6:8091 --group-name East \ --server-username Administrator --server-password password
man:couchbase-cli-rebalance[1], man:couchbase-cli-recovery[1], man:couchbase-cli-server-add[1]