We have a single node Elastic Search cluster (v7.17.1) running. We are creating a new multi node Master-Data-Client Elastic Search (v7.17.1) cluster for high-availability.
We are using Kubernetes and helm charts for deploying these clusters.
Unfortunately we can keep either the single node cluster or the new multi node cluster running at a given point in time.
We were able to take a snapshot of the single node cluster on a FS (in a directory inside the single node cluster). We then copied the snapshot contents from this directory and placed it in a directory outside the single node cluster. Then we shutdown the single node cluster.
We then installed the multi node cluster and now want to restore the snapshot copied/saved in the outside directory; as explained above.
How can this be achieved?
Or is there any other way where we could migrate data from Single Node to Multi Node Cluster; but with the limitation that only one of the cluster (existing single node or new multi node) could be running at a given point in time?
Did you take this using the snapshot API in Elasticsearch?
You should be able to move the snapshot to shared storage, mount it to all the nodes in the cluster and then use the restore API.
If you do not have shared storage available you may be able to copy the snapshot into all nodes and mount the repository as read-only with verify set to false.
Yes we used "PUT /_snapshot/{repository}" and "PUT /_snapshot/{repository}/" to create a repository and get a snapshot when the single node cluster was running.
How can this be done? Can you please provide few links?
We have created a new directory (ESbackup) inside the new Multi Node Cluster and manually copied the snapshot .dat, .snap and all the other files/directories that were created when the snapshot was created in the single node cluster. This new directory(ESbackup) was mentioned under path.repo param inside the elasticsearch.yaml file of the new Multi Node Cluster. Is there anyway to restore these copied snapshot files in the new Multi Node Cluster? If yes can you please let us know how?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.