Elasticsearch 8.3.2 docker org.elasticsearch.gateway.CorruptStateException

Installed ES 8.3.2 using docker on my local dev machine. Before this installation I had ES 6.3.2 running and I had deleted all my test indices in 6.3.2 before installing 8.3.2 so that I would avoid any v6 issues with 8 on my local dev. When I attempt docker-compose up --build I get this error:

es_usamm-db          | org.elasticsearch.ElasticsearchException: failed to bind service
es_usamm-db          | Likely root cause: org.elasticsearch.gateway.CorruptStateException: Format version is not supported. Upgrading to [8.3.2] is only supported from version [7.17.0].

How do I completely wipe out any elasticsearch docker data so that I can just install 8.3.2? I thought deleting all indices in 6.3.2 would have allowed a clean install of 8.3.2, but there must still be some 6.3.2 data persisting somewhere?

Here is my docker-compose.yml

services:
  elasticsearch:
    container_name: es_usamm-db
    image: docker.elastic.co/elasticsearch/elasticsearch:8.3.2
    environment:
      - node.name=es01
      - cluster.name=es-docker-cluster
      - bootstrap.memory_lock=true
      - "ES_JAVA_OPTS=-Xms512m -Xmx512m"
      - xpack.security.enabled=false
      - discovery.type=single-node
    ulimits:
      memlock:
        soft: -1
        hard: -1
    volumes:
      - data01:/usr/share/elasticsearch/data
      - ./server/elastic/scripts:/server/elastic/scripts
    ports:
      - 9200:9200
    networks:
      - default
  kibana:
    container_name: kibana_usamm-db
    image: docker.elastic.co/kibana/kibana:8.3.2
    ports:
      - 5601:5601
    networks:
      - default

Can you elaborate more on what you did when you said you deleted the indices?

You should really be using a different volumes: value, if data01 was what you used for 6.X. That way you don't need to worry about deleting anything, as it'll be a new data directory.

I did end up using a different volume i.e. data02 to get past the error.

To answer your question, when v6 was running, in dev tools kibana, I ran DELETE _all. Everything was wiped out, which was confusing when I tried to run v8 on the data01 volume. Anyway, changing the volume fixed it.

1 Like

That's what I figured you did.

But that doesn't remove things like cluster state, which is probably part of the reason you ran into this issue.

2 Likes