OVH as hosted elasticsearch

"Cold" ou "Cloud"? Parce que j'étais parti dans le surgelé, là :smile:

Ah ok, ben je n'ai pas trouvé avec le moteur de recherche, merci pour le lien.

> Si un fragment manquant est détecté, il est automatiquement reconstruit via un algorithme spécifique utilisant les autres fragments.

Ca , je ne vois pas bien comment ça peut marcher, y'a une boule de cristal dans la machine ? :smile:

@CM63 The data is hosted fragmented across many disks.

There is always some sort of redundancy and control codes that allows reconstructing the data when a disk is lost.

Basically there is always at least either:
- another copy of the data
- Something you can calculate the data from of you know parts of it doing a lot of calculations

RAIDs works this way too.




----------


@PierreD Is any preliminary information available ?

So it is redondancy rather than using other part of the data, so I aggree.

maybe @PierreD i'm trying to understand. Does it works like this?

* You buy a certain number of indices
* each indice can have 1 to 10 shards
* if you want more than 25gb per indice you buy extra shards for that indice

What does LDP mean?
What do you mean by correlating logs with static content?

@PierreD ok. I think that operation is often called "fill", when you have a database and fill the response of one query with extra data from another query.

Anyhow that may very well be what we're looking for.
Will it be a shared cluster or will you run the clusters in isolated containers/vm and provide a small cluster with an X amount of resources to each user?

@PierreD qbox.io have a series of articles on all the problems they had doing a shared ES cluster and that ended up i developing this software https://supergiant.io/

@PierreD Are there any updates on the availability of elasticsearch as a service?

@PierreD vague