Installing more than one replica doesn't work about meilisearch-kubernetes HOT 8 CLOSED

Hi @laurenz-glueck,
Thanks for asking. The team has planned to move forward on the High Availability around the end of the year but it is possible that it will change. Nevertheless, it is still a goal that we would really like to achieve.

from meilisearch-kubernetes.

Hey @deshetti,

Indeed, this seems to be an issue, but as for now MeiliSearch itself wouldn't support it either, since it desn't allow reading from the same database by several instances. The core team is workin on a high availability and consensus feature that will let us create several instances for replication and consistency. This should be available in the next few months (can't tell what would be the eta exactly), but as it should be a consensus algorythm instances would own their own isolated storage and at the same time data would be syncronised in different instances of a cluster, fix this problem (I think). WDYT?

from meilisearch-kubernetes.

Hi guys! I'm also expecting a real-time synchronization feature for mutiple databases. This would be a lot better to use in production than using snapshots/dumps.

from meilisearch-kubernetes.

Hi @simonxxz

This is in MeiliSearch's roadmap indeed, but as for now, the engine is not capable of providing replication and a distributed deployment. If this feature is important for you I invite you to give it a little upvote in the roadmap, but be reassured that the team is working on that :)

High Availability on our roadmap

from meilisearch-kubernetes.

Is there any update on this topic or an eta?

from meilisearch-kubernetes.

Closing due to lack of answer, feel free to re-open if needed of course 🙂

from meilisearch-kubernetes.

Information update:
For now, Meilisearch doesn't manage high availability, and for the moment the team doesn't work on it anymore. I close this topic until the subject is put back in the roadmap.

from meilisearch-kubernetes.

..

from meilisearch-kubernetes.