Restart Elastic search without data loss single node cluster

I want to restart my Elasticsearch service which is running on EC2, data stored on EFS mounted on EC2. As in current yml file there is no security related configurations present, I want add these and restart Elasticsearch service. Whether it will impact existing data, I'm using single node cluster . It is a production env.

Please help!

A one-node cluster will recover from a restart as long as the data on disk persists across the restart. Of course it will be unavailable while it’s restarting.

I recall issues using EFS for data storage with Elasticsearch, so would recommend switching to EBS.

Performance issues, not data loss tho. For light workloads it’s probably fine.

I vaguely recall EFS not behaving as a file system the way Elasticsearch requires. In the docs it still states the following:

Historically, shared network filesystems such as EFS have not always offered precisely the behaviour that Elasticsearch requires of its filesystem, and this has been known to lead to index corruption.

Is that no longer true?

Hmm not sure these days, I haven’t heard of anyone encountering such an issue for ages. But maybe that’s because not many people use EFS for this purpose rather than because they fixed the bugs.

1 Like