Elasticsearch "stuck" indices

Hello,

we have a problem with a few indicies which are not "moveable".
We have a ILM with a warm and cold phase.
Warm phase:
After 48 hours, shrinks and force merge

Cold phase:
After 17 days
Freeze and move to other data nodes

Now we saw that we're getting ILM errors on the warm phase, because we have 5 shards per indicies and he can't shrink/merge with an odd number of shards. We fixed the template, so we wont have the issue in future.

Anyways, the other Indicies are now stuck. They're in ILM warm phase, but they arent merged or shrinked. But they missing all replicas and are not "moveable". We have to move them to the other servers because the storage is getting low on our hot nodes.

But we're not able to do so. We tried:

  • Close and open the index
  • Remove ILM policy
  • Allocate manually to other server
    But they're always started on the hot nodes and won't go to out warm nodes.

Do you have any tips / hints for us? Thanks!

Maybe reindex old to new, then delete old.

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