OpenSearch instance change hangs on copying shards

Question

Good morning everybody,
I have a test OpenSearch domain running on a t3.medium instance. After creating an index, the cluster status changed on red.
To solve, I tried to changes the instance to a superior one, but the dry run told me that was not possible. Hence, I changed it to another t3.medium.

The migration started, but the process is stuck at step 4 of 5, copying the shard 19 of 41. It's been there for 2 days. 
In the while, I managed to delete the incriminating index, so the cluster status reverted to green. Nevertheless, the migration is still hanging.
Any help, please?

Accepted Answer

Hi, when the cluster is RED, any blue green deployments that you trigger as a result of configuration changes will get stuck in processing state. Its advised to make sure the cluster is either is green or yellow state before you make any changes to the cluster. Managed Opensearch service performs multiple auto retries (before it gives up) to see if the cluster is out of RED state so that it can finish the Blue Green deployment. Now that the cluster is stuck in processing, there is not much you could do but open a support case for the AWS teams to manually clear the stuck state and continue with blue green deployment.

OpenSearch instance change hangs on copying shards

Relevant content