Migrating previous generation nodes
Previous generation nodes are node types that are being phased out. If you have no existing clusters using a previous generation node type, ElastiCache does not support the creation of new clusters with that node type.
Due to the limited amount of previous generation node types, we cannot guarantee a successful replacement when a node becomes unhealthy in your cluster(s). In such a scenario, your cluster availability may be negatively impacted.
We recommend that you migrate your cluster(s) to a new node type for better
availability and performance. For a recommended node type to migrate to, see Upgrade Paths
Migrating nodes on a Redis OSS cluster
The following procedure describes how to migrate your Redis OSS cluster node type using the ElastiCache Console. During this process, your Redis OSS cluster will continue to serve requests with minimal downtime. Depending on your cluster configuration you may see the following downtimes. The following are estimates and may differ based on your specific configurations:
-
Cluster mode disabled (single node) may see approximately 60 seconds, primarily due to DNS propagation.
-
Cluster mode disabled (with replica node) may see approximately 1 second for clusters running Redis OSS 5.0.6 and above. All lower version can experience approximately 10 seconds.
-
Cluster mode enabled may see approximately 1 second.
To modify a Redis OSS cluster node type using the console:
-
Sign in to the Console and open the ElastiCache console at https://console.aws.amazon.com/elasticache/
. -
From the navigation pane, choose Redis OSS clusters.
-
From the list of clusters, choose the cluster you want to migrate.
-
Choose Actions and then choose Modify.
-
Choose the new node type from the node type list.
-
If you want to perform the migration process right away, choose Apply immediately. If Apply immediately is not chosen, the migration process is performed during the cluster's next maintenance window.
-
Choose Modify. If you chose Apply immediately in the previous step, the cluster's status changes to modifying. When the status changes to available, the modification is complete and you can begin using the new cluster.
To modify a Redis OSS cluster node type using the Amazon CLI:
Use the modify-replication-group API as shown following:
For Linux, OS X, or Unix:
aws elasticache modify-replication-group / --replication-group-id
my-replication-group
/ --cache-node-typenew-node-type
/ --apply-immediately
For Windows:
aws elasticache modify-replication-group ^ --replication-group-id
my-replication-group
^ --cache-node-typenew-node-type
^ --apply-immediately
In this scenario, the value of new-node-type
is the
node type you are migrating to. By passing the --apply-immediately
parameter, the update will be applied immediately when the replication group
transitions from modifying to available
status. If Apply immediately is not chosen, the migration
process is performed during the cluster's next maintenance window.
Note
If you are unable to modify the cluster with an
InvalidCacheClusterState
error, you need to remove a
restore-failed node first.
Fixing or removing restore-failed-node(s)
The following procedure describes how to fix or remove restore-failed node(s) from your Redis OSS cluster. To learn more on how ElastiCache node(s) enter a restore-failed state, see Viewing ElastiCache Node Status. We recommend first removing any nodes in a restore-failed state, then migrating the remaining previous generation nodes in the ElastiCache cluster to a newer generation node type, and finally adding back the required number of nodes.
To remove restore-failed node (console):
-
Sign in to the Console and open the ElastiCache console at https://console.aws.amazon.com/elasticache/
. -
From the navigation pane, choose Redis OSS clusters.
-
From the list of clusters, choose the cluster you want to remove a node from.
-
From the list of shards, choose the shard you want to remove a node from. Skip this step if cluster mode is disabled for the cluster.
-
From the list of nodes, choose the node with a status of
restore-failed
. -
Choose Actions and then choose Delete node.
Once you remove the restore-failed node(s) from your ElastiCache cluster, you can now migrate to a newer generation type. For more information, see above on Migrating nodes on a Redis OSS cluster.
To add back nodes to your ElastiCache cluster, see Adding nodes to an ElastiCache cluster.