how to restore elasticsearch indices from S3 to bl

问题:

I have an S3 bucket of elasticsearch snapshots created by a curator job. I want to be able to restore these indexes to a fresh cluster using the S3 bucket. The target elasticsearch cluster does not have access to the source elasticsearch cluster by design.

I've installed cloud-aws plugin on the es client for the target cluster and I set permissions to the S3 bucket using environment variables. I have the config and action file in place for curator. I've verified the AWS permissions to the S3 bucket, but I'm not sure how to verify the permissions from the elasticsearch cluster's perspective. When I try running the curator job I get the following:

get_repository:662 Repository my-elk-snapshots not found.

I know that if I were to use elasticsearch directly I would need to create a reference to the S3 bucket so that the cluster knows about it. Is this the case for a fresh restore? I think that curator uses the elasticsearch cluster under the hood, but I'm confused about this scenario since the cluster is essentially blank.

回答1:

How did you add the repository to the original (source) cluster? You need to use the exact same steps to add the repository to the new (target) cluster. Only then will the repository be readable by the new cluster. That's why you're getting the "repository not found" message. It has to be added to the new cluster so that snapshots are visible, and therefore able to be restored.