Elasticsearch Closing Index after removing read-only index block followed by Restart

Elastic Stack Elasticsearch
#1

Hi,

We are on Elasticsearch 6.0 and we need help on the following scenario:

  1. Index A is in "Open" state
  2. Elasticsearch crossed flood_stage watermark, Index A got read-only index block
  3. Resolve flood_stage watermark, Remove read-only index block from index A by setting
    {
      "index.blocks.read_only_allow_delete": null
    }
  1. Index A is still in "Open" state
  2. Restart Elasticsearch
  3. Index A is closed when Elasticsearch is back online

We want to confirm whether this is the expected behavior, and if there is any way to avoid this auto-close behavior.

Thank you!

#2

You really should upgrade as a matter of urgency. That version is ancient, it reached the end of its supported life 18 months ago. In more recent versions the read_only_allow_delete block is removed automatically.

I don't think you are describing what you saw accurately. Elasticsearch doesn't auto-close indices. What makes you think it was closed?

#3

Yes, migration plan is ongoing :slight_smile:

Awesome!

Well, upon ES startup, the index is in "Closed" state.
I guess for now we will just avoid doing Elasticsearch restart and watch closely on this behavior until we upgraded to ES7.

Thanks for the input!

#4

Where are you seeing this, though? Can you share the API request that gives you this information, and the response?

#5

I'm using cat API, i.e. GET /_cat/indices

#6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.