Chris Jensen

The stream approach should work I think (given limited size streams), but would require that any rpc library to implement that kind of processing. Is there any way to do...

@changetoblow I worked around it via the loopback address, unfortunately...

After rewriting the IO layer, this is no longer an issue for me specifically. What was occurring was that if client requests were coming in they would each start a...

This should be closed since mininet now allows for float values: https://github.com/mininet/mininet/blob/6eb8973c0bfd13c25c244a3871130c5e36b5fbd7/mininet/link.py#L303C27-L303C27

@ahrtr Just checking my intuition here. Given there is a lagging node that calls and fails to be elected, you want to bias the retried election to the other replicas?...

> @erikgrinaker @pav-kv PTAL when you have time. > > @Cjen1 do you have any immediate concern? > > Thanks. I am still concerned about this impacting the minimal viable...

I've run this PR through [reckon](https://github.com/cjen1/reckon), using 100 leader loss events for each config, varying the number of nodes (3 and 5) and the election timeout.  I think the...

@pavelkalinnikov Yep this is exactly where the idea came from (I was doing some comparisons of leader election performance between multipaxos and raft). Also that sounds like a good idea...

> It doesn't do exponential backoff, but each node's election timeout is randomized by a factor of [1-2) to reduce the chance of ties: Ah! Then randomisation of the terms...