Is it possible to replicate the index WITHOUT yellow status? (replication disabled due to faster reindex)

Hi there,

due to speedup of re-indexing, I'm setting "number_of_replicas": 0 while re-indexing is running.
Once finished, than I'm setting 1 replica:

PUT bmpf_claim_docs_v2/_settings
    "number_of_replicas": 1

Until the replicas will be created and distributed, ES cluster is in yellow state, it will create the ticket etc.

So my question is: Is it possible to replicate index with some advanced command to avoid yellow state? (still being able to re-index same speed like when replicas set to 0)

Note: with replicas indexing speed ~1000/s, without 1500/s (re-indexing in same cluster, 13M docs, 500GB primary, 15shards, cluster has 15 data nodes)

That's the expected behaviour, and no you cannot bypass it. Yellow health is expected sometimes and will normally resolve to green automatically, there's no need to create a ticket when it happens.

ticket in case of "yellow state" is created automatically as part of monitoring for similar cases like that:

Maybe we should monitor only "red status".

Anyway, thanks for the quick answer

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