Hello everyone,
I am new to the forum and basically self-taught, so I apologize in advance if I lack some basic knowledge. Thank you for your understanding.
The elastic cluster works on a k8s cluster as pods, but reading the documentation, I cannot understand clearly three things:
1- Do frozen nodes require persistent storage, or are they okay with volatile storage? Could I freely destroy a frozen pod without any problems and have it reappear elsewhere? Or would it lose the previously known frozen indexes?
2- I thought that frozen indexes (managed by ILM, not manually mounted) were a separate replica from the snapshots I take daily. Is that correct? Or are they based on old snapshots? (I have both on the same repository.)
3- (derived from point 2...) If I delete all daily snapshots from the repository, will the frozen indexes managed by ILM remain safe?
EDIT: Small clarifications that may be useful:
- The daily snapshots are created with a policy in Stack Management → Snapshot and Restore → Policies.
- The frozen indices are managed via ILM. In fact, in the snapshot repository I can see one snapshot per index, with the ILM policy name included. Since many indices have moved to the frozen tier, the remote storage usage has increased by about 10 TB.