DRAFT: Use new RA callback

[SimonUnge]

Proposed Changes

Draft for @kjnilsson

Types of Changes

What types of changes does your code introduce to this project?
Put an x in the boxes that apply

• Bug fix (non-breaking change which fixes issue #NNNN)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause an observable behavior change in existing systems)
• Documentation improvements (corrections, new content, etc)
• Cosmetic change (whitespace, formatting, etc)
• Build system and/or CI

Checklist

Put an x in the boxes that apply.
You can also fill these out after creating the PR.
If you're unsure about any of them, don't hesitate to ask on the mailing list.
We're here to help!
This is simply a reminder of what we are going to look for before merging your code.

• I have read the CONTRIBUTING.md document
• I have signed the CA (see https://cla.pivotal.io/sign/rabbitmq)
• I have added tests that prove my fix is effective or that my feature works
• All tests pass locally with my changes
• If relevant, I have added necessary documentation to https://github.com/rabbitmq/rabbitmq-website
• If relevant, I have added this change to the first version(s) in release-notes that I expect to introduce it

Further Comments

If this is a relatively large or complex change, kick off the discussion by explaining why you chose the solution you did and what alternatives you considered, etc.

[kjnilsson]

I think the general approach is pretty much what we designed. I am concerned that there are functions called in eval_members that are potentially blocking such as rabbit_nodes:list_running() and we'd have to see if it feasible to run all "grow" logic in a non blocking way to avoid locking the Ra leader process up.

That said this is very much on the right track and we just need to investigate a bit further if the logic can be made safe to run inside the ra process.

[SimonUnge]

I think the general approach is pretty much what we designed. I am concerned that there are functions called in eval_members that are potentially blocking such as rabbit_nodes:list_running() and we'd have to see if it feasible to run all "grow" logic in a non blocking way to avoid locking the Ra leader process up.

That said this is very much on the right track and we just need to investigate a bit further if the logic can be made safe to run inside the ra process.

Awesome. Any hints on how to investigate?

I was thinking of somehow call the 'eval_members' in its own process, and somehow get the result back to the RA process to run the grow code.

[SimonUnge]

@kjnilsson Perhaps the eval_member callback could provide an optional 'reply fun', and the implementer could use that fun to report the result to the instead of directly responding, i.e spawn a process that does the eval, and add the 'effects' to the fun, and the funs job would be to send a msg back to the ra_server_proc leader process?