I've got an ES database that I accidentally upgraded to 7.16 following a stupid apt-get update
For it to remain compatible with my applications, I need to roll it back to 7.10
There doesn't seem to be any offical advice on how to do this, but surely the upgrade cannot be one way. I'll take any route, no matter how convoluted and difficult.
Please please is there a way to get my data back into ES 7.10?
Elasticsearch does not support downgrades. The only way to get back to a pervious version is to spin up an empty cluster and restore a snapshot taken before the upgrade. If you do not have a snapshot I believe you will need to reindex from remote into a new empty 7.10 cluster.
So I can see there is some documentation on reindex from remote here:
How does this work with multi-node setups? (is there anything closer to a runbook/operational guide to how to execute this?)
Does the new cluster need to have the same number of nodes, or can it just have a single node with sufficient storage to hold all data from all nodes?
Assume that I don't have the same hardware available to spin up a replacement cluster wholesale, and instead need to use the same hardware to host the restored 7.10 nodes.
Once the data is moved to the new cluster, is it just a matter of reindex from remote back?
How long approximately does reindex from remote take? Eg. is there a way I can estimate downtime based on gb of data?
Not that I know of. You need to run it index by index.
Cluster size does not need to be the same.
Spin up a single 7.10 node if this can hold the data. Once reindexing is complete you can add remove the 7.16 nodes and create new 7.10 nodes that you add to the 7.10 node to form a cluster.
It depends on the data, process and hardware so is impossible to tell.
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.