Index xxxxx belongs to a version of Kibana

I'm using docker-kibana and am getting the message:

FATAL Error: Index mixpanel belongs to a version of Kibana that can not be automatically migrated. Reset it or use the X-Pack upgrade assistant.

Followed by exit code 1.

I searched a little for this error and it seems that the problem is that the version of the index does not correspond with the version of kibana, but then I was in a doubt. I'm running the kibana 6.5.4 image which is the same version that the index was created in the elastic search.

xxxxxxx / _settings? pretty

{
   "xxxxx": {
     "settings": {
       "index": {
         "number_of_shards": "5",
         "provided_name": "xxxxx",
         "max_result_window": "20000",
         "creation_date": "1548417754749",
         "number_of_replicas": "1",
         "uuid": "_NFW-GcoQyGhjwA49Ca59g",
         "version": {
           "created": "6050499"
         }
       }
     }
   }
}

Could someone help me with this?

Hi, is there a docker volume for Elasticsearch data that was mounted in a container for an earlier version (like 5.x) of the Elasticsearch image?

Can you clarify if mixpanel is the name of Kibana's index? (The one that is .kibana by default).

If you start ES and Kibana 5.x, then re-use the data directory to start ES and Kibana 6, the upgrade migration assistant finds that you need to update your Kibana index (which seems to be called mixpanel) first. See https://www.elastic.co/guide/en/kibana/current/migrating-6.0-index.html

Yes, mixpanel is the index.

That response about upgrading to kibana 6 I had found but before I started trying I wanted to understand better. See, the ES is already in version 6 as shown below:

{
  "name": "instance-0000000000",
  "cluster_name": "0324510767eec4ff3a42270418a9ffd6",
  "cluster_uuid": "fM6HHfqDSjuBT7sXR4DnAw",
  "version": {
    "number": "6.5.4",
    "build_flavor": "default",
    "build_type": "tar",
    "build_hash": "d2ef93d",
    "build_date": "2018-12-17T21: 17: 40.758843Z",
    "build_snapshot": false,
    "lucene_version": "7.5.0",
    "minimum_wire_compatibility_version": "5.6.0",
    "minimum_index_compatibility_version": "5.0.0"
  },
  "tagline": "You Know, for Search"
}

The kibana image I'm using is also 6.5.4 as shown by me dcker-compose.yml in the first post.

That being said, why would I need to upgrade to version 6, since I already have both systems in 6.5.4?

Another detail, I have a kibana that was built through an AWS tool that connects to this same Elastic Search that I am trying to connect in the local image and from what I could see the kibana in AWS is using version 6.5.4 as well and it works. I checked the kibana version on AWS through the tag <kbn-injected-metadata> that I found by parsing the HTML source code.

Is this clear in my doubts?

Your Kibana index was created in 5.x, back when Kibana's data model for that index was a little different. How it organized your saved objects was different.

When you retained that index after upgrading to 6.x, Kibana finds that old data model and asks you to go back to v5.6 first, run the upgrade tool on that index, and then try again in 6.x. Kibana 5.6 can work with the old data model, and the new data model.

In other words, when upgrading from 5 to 6, a migration step was skipped, so you'll have to backtrack a bit before you can have a successful upgrade.

I think I understand the problem, thank you!

1 Like

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