[Kibana 7.17.1] Index patterns disappear after restoring VM snaphots of ES cluster machines

We're running an ES test cluster of 3 machines with the following nodes:

  1. master, coordinating only, data
  2. master, coordinating only, data
  3. master, ingest, data

We have three index types in this cluster:

  1. shop-<YYYY.MM.DD>
  2. k8s-nginx-ingress-<YYYY.MM.DD>
  3. k8s-test-<YYYY.MM.DD>

each of them, you probably guessed it, having one index created daily.

If I restore all three machines one by one (and wait between them until the reorganization of shards has finished) I can see 46 indices in our Cerebro across all three types (three are from today, Sep. 1st, created newly).

But, when logging in to Kibana with the elastic user > Default space I only see an index pattern for shop-*. We (me and a few colleagues) are pretty sure that there were index patterns for all three types before.

What could possibly have gone wrong there?

P.S.: This appeared in the process of upgrading to ES/Kibana 8.y.z and we will not upgrade our prod cluster unless it works properly in the test cluster. We have 2,865 indices, 5,371 shards, 3,761,281,899 docs, 2.25 TB in our prod cluster, so we want to be absolutely sure what's going on before upgrading it.

See:

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