Replication factor

[madshargreave]

Would it be possible to add a replication factor? Ie. only replicate this table on N nodes in the entire cluseter

[beardedeagle]

Like at random or do you want to control where they land?

[beardedeagle]

I am thinking of these two scenarios:

• I want this table to only exist in N nodes
• I want this table to only exist on these specific nodes

And I think I have an idea of how to implement both in one go.

[madshargreave]

Exactly the above :)

[beardedeagle]

@madshargreave I've been thinking about this some more and I'm unsure if I want this table to only exist on these specific nodes is a concern of the library. I say this for a couple of reasons:

• If you are using something like libcluster, then you should be making use of it's whitelist and blacklist capabilities
• If you are manually managing membership then you have control of this anyway.

Now I could still be convinced of this if the usecase makes sense, but as of the moment I don't think it's a concern of Mnesiac.

I want this table to only exist in N nodes has some use, though this type of check would only happen at startup.

[beardedeagle]

I've begun work on this in the replication_factor branch for at least a partial implementation of this request.

[beardedeagle]

as a heads up, @astutecat suggested also adding a 3rd option for replicating by percentage, so:

• I want this table to only exist in N nodes
• I want this table to only exist on these specific nodes
• I want this table to exist on N% of nodes

[beardedeagle]

Sorry, I've been away for some time. I'm getting back to this. Apologies for the inconvenience.