Troubleshoot Cluster issues

Available in Classic and VPC

You might run into the following problems when using Search Engine Service. Find out causes and possible solutions.

Cluster status - warning

Cluster status shows Warning.

Cause

A warning appears when one or more indexes are recovering during rebalancing after a restart.

Solution

• See Cluster status > Warning.
• Check the status of the affected index or indexes. If index sizes are large, recovery after restart can take from minutes to hours. For an index of about 400 GB, recovery can take 2-3 hours or longer.
• If the warning is yellow, the issue affects only part of the cluster. The cluster continues operating, but I/O may be unavailable for the affected index. For managing indexes in an OpenSearch cluster, operate according to the official documentation on opensearch.org.

Change redundancy settings for the manager node

I want to change redundancy settings for the manager node while the cluster is running.
After creating a manager node, you cannot separate data nodes and master nodes while in operation.

Cause

You cannot change manager‑node redundancy or split data and master nodes while a cluster is running.

Solution

• As explained in Cause, you cannot split data and master nodes after cluster creation. Decide redundancy settings carefully when creating the cluster.
• Because you cannot change redundancy or node roles on a running cluster, create a new cluster if you need different settings.

Cannot create new clusters on G1 (first‑generation) servers (Classic)

In the Classic environment, you cannot create new clusters on G1 (first-generation) servers.

Cause

Starting April 22, 2022, in the Classic environment you can create new clusters only on Classic G2 (second‑generation) servers.

Solution

As explained in the cause section, it is a normal operation and it does not need any action. Create clusters on Classic G2 (second‑generation) servers.