Maintenance Mode - Redis

Configures [quorum-only mode]({{< relref "/operate/rs/7.4/clusters/maintenance-mode#activate-maintenance-mode" >}}) on a node.

`node maintenance_mode on`

Migrates shards out of the node and turns the node into a quorum node to prevent shards from returning to it.

rladmin node <ID> maintenance_mode on
        [ keep_slave_shards ]
        [ evict_ha_replica { enabled | disabled } ]
        [ evict_active_active_replica { enabled | disabled } ]
        [ evict_dbs <list of database names or IDs> ]
        [ demote_node ]
        [ overwrite_snapshot ]
        [ max_concurrent_actions <integer> ]

Parameters

Parameter	Type/Value	Description
node	integer	Turns the specified node into a quorum node
demote_node		If the node is a primary node, changes the node to replica
evict_ha_replica	`enabled`
`disabled`	Migrates the HA replica shards in the node
evict_active_active_replica	`enabled`
`disabled`	Migrates the Active-Active replica shards in the node
evict_dbs	list of database names or IDs	Specify databases whose shards should be evicted from the node when entering maintenance mode.

Examples: $ rladmin node 1 maintenance_mode on evict_dbs db:1 db:2 $ rladmin node 1 maintenance_mode on evict_dbs db_name1 db_name2 | | keep_slave_shards | | Keeps replica shards in the node and demotes primary shards to replicas.

Deprecated as of Redis Enterprise Software 7.4.2. Use evict_ha_replica disabled evict_active_active_replica disabled instead. | | max_concurrent_actions | integer | Maximum number of concurrent actions during node maintenance | | overwrite_snapshot | | Overwrites the latest existing node snapshot taken when enabling maintenance mode |

Returns

Returns OK if the node was converted successfully. If the cluster does not have enough resources to migrate the shards, the process returns a warning.

Use [rladmin status nodes]({{< relref "/operate/rs/7.4/references/cli-utilities/rladmin/status#status-nodes" >}}) to verify the node became a quorum node.

Example

$ rladmin node 2 maintenance_mode on overwrite_snapshot
Found snapshot from 2024-01-06T11:36:47Z, overwriting the snapshot
Performing maintenance_on action on node:2: 0%
created snapshot NodeSnapshot<name=maintenance_mode_2024-01-11_20-25-37,time=None,node_uid=2>

node:2 will not accept any more shards
Performing maintenance_on action on node:2: 100%
OK
$ rladmin status nodes
CLUSTER NODES:
NODE:ID ROLE   ADDRESS    EXTERNAL_ADDRESS  HOSTNAME     SHARDS CORES       FREE_RAM         PROVISIONAL_RAM  VERSION   STATUS
*node:1 master 192.0.2.12 198.51.100.1      3d99db1fdf4b 5/100  6           14.21GB/19.54GB  10.62GB/16.02GB  6.2.12-37 OK
node:2  slave  192.0.2.13 198.51.100.2      fc7a3d332458 0/0    6           14.21GB/19.54GB  0KB/0KB          6.2.12-37 OK
node:4  slave  192.0.2.14                   6d754fe12cb9 5/100  6           14.21GB/19.54GB  10.62GB/16.02GB  6.2.12-37 OK

`node maintenance_mode off`

Turns maintenance mode off and returns the node to its previous state.

rladmin node <ID> maintenance_mode off
        [ { snapshot_name <name> | skip_shards_restore } ]
        [ max_concurrent_actions <integer> ]

Parameters

Parameter	Type/Value	Description
node	integer	Restores the node back to the previous state
max_concurrent_actions	integer	Maximum number of concurrent actions during node maintenance
skip_shards_restore		Does not restore shards back to the node
snapshot_name	string	Restores the node back to a state stored in the specified snapshot

Returns

Returns OK if the node was restored successfully.

Use [rladmin status nodes]({{< relref "/operate/rs/7.4/references/cli-utilities/rladmin/status#status-nodes" >}}) to verify the node was restored.

Example

$ rladmin node 2 maintenance_mode off
Performing maintenance_off action on node:2: 0%
Found snapshot: NodeSnapshot<name=maintenance_mode_2022-05-12_20-25-37,time=2022-05-12T20:25:37Z,node_uid=2>
Performing maintenance_off action on node:2: 0%
migrate redis:12 to node:2: executing
Performing maintenance_off action on node:2: 0%
migrate redis:12 to node:2: finished
Performing maintenance_off action on node:2: 0%
migrate redis:17 to node:2: executing

migrate redis:15 to node:2: executing
Performing maintenance_off action on node:2: 0%
migrate redis:17 to node:2: finished

migrate redis:15 to node:2: finished
Performing maintenance_off action on node:2: 0%
failover redis:16: executing

failover redis:14: executing
Performing maintenance_off action on node:2: 0%
failover redis:16: finished

failover redis:14: finished
Performing maintenance_off action on node:2: 0%
failover redis:18: executing
Performing maintenance_off action on node:2: 0%
failover redis:18: finished

migrate redis:21 to node:2: executing

migrate redis:19 to node:2: executing
Performing maintenance_off action on node:2: 0%
migrate redis:21 to node:2: finished

migrate redis:19 to node:2: finished

failover redis:20: executing
Performing maintenance_off action on node:2: 0%
failover redis:20: finished
Performing maintenance_off action on node:2: 0%
rebind endpoint:6:1: executing
Performing maintenance_off action on node:2: 0%
rebind endpoint:6:1: finished
Performing maintenance_off action on node:2: 100%
OK
$ rladmin status nodes
CLUSTER NODES:
NODE:ID ROLE   ADDRESS    EXTERNAL_ADDRESS  HOSTNAME      SHARDS CORES       FREE_RAM        PROVISIONAL_RAM  VERSION   STATUS
*node:1 master 192.0.2.12 198.51.100.1      3d99db1fdf4b  5/100  6           14.2GB/19.54GB  10.61GB/16.02GB  6.2.12-37 OK    
node:2  slave  192.0.2.13 198.51.100.2      fc7a3d332458  5/100  6           14.2GB/19.54GB  10.61GB/16.02GB  6.2.12-37 OK    
node:4  slave  192.0.2.14                   6d754fe12cb9  0/100  6           14.2GB/19.54GB  10.69GB/16.02GB  6.2.12-37 OK