Execution order of multiple indexes during _reindex call

Elastic Stack Elasticsearch
1

Hi all :wave:,

During _reindex call, if I specify remote source from multiple indexes, In which order are those indexes going to be processed?

I saw in documentation that there was sort option in reindex but it is deprecated since 7.6, and that sort was never guaranteed.

In case above was not clear enough, I'm trying to do something like this:

POST _reindex/?wait_for_completion=false&wait_for_active_shards=all
{
  "source": {
    "remote": {
      "host": "https://example.com:443"
    },
    "index": "my-index-*",
    "size": 5000,
    "query": {
      "match_all": {}
    }
  },
  "dest": {
    "index": "my-index-new"
  }
}

And reason why I am asking is because I have quite some duplicate document ids spread-ed over my-index-yyyy-MM-dd on remote, not all have desired states recorded, but the latest inserted version is the one I want to keep on my my-index-new, and it "seems" that documents are kinda updated in that order.

So maybe coincidence but from what I am seeing based on results, it looks like execution is done based on timestamp/insertion order of remote indexes? Or there is no guarantee for this?

remote es version: 6.6
destination es version: 7.8

1 Like
2

Anyone has some pointers for this one?

3

Answer: execution is not done based on timestamp/insertion order of remote indexes, few checks based on above assumption were coincidence

4

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