Optimizing Elasticsearch Snapshot Recovery for Node Disk Space Utilization

[jakub0011 (Jakub)]

I'm seeking advice on optimizing the snapshot recovery process in our Elasticsearch cluster, which consists of 8 nodes. Currently, when recovering various snapshots, the indices are restored to nodes based on the percentage of disk space utilization in the "/var/lib/elasticsearch" directory.

I would like to configure the system so that indices are distributed to nodes where the disk space utilization is less than, for example, 70%. At present, the recovery process initiates on nodes where disk space utilization is as high as 89%. My goal is to ensure that the snapshot recovery occurs on nodes with a disk space utilization below the specified threshold of 70%.

Can anyone guide me on how to achieve this configuration or share insights on best practices for optimizing snapshot recovery based on disk space utilization?

Thank you in advance for your assistance!

elasticsearch version: 7.10

[DavidTurner]

Hi there,

@jakub0011

elasticsearch version: 7.10

This version is ancient, you need to upgrade as a matter of some urgency.

Snapshot restores are allocated just like other shards, so you can control the process by adjusting things like the disk watermarks.