Madelyn Olson

>Also, as the benchmarking tool is single threaded sometimes running one benchmark process is not enough to saturate the server. Hence, we end up running multiple benchmark process. With that...

I like this idea! I've seen issues in the past, especially with cluster mode, where a node ends up being marked as a primary with no data. It seems like...

> I'd think when replication is reconfigured to change which nodes are primaries and which are replicas that would flush out the old primary state data that could wedge up...

> The core problem is that the primary didn't stay "down" long enough to trigger the election reliably. This is an unsolvable problem since we prioritize availability over consistency, and...

[Deleted] I just realized the case I was mentioning only applies to how AWS does it, so that's just a problem that we have. I think Ping is right that...

@valkey-io/core-team I guess let's vote on all of these. This sounds like a good change to me.

It might make sense to have a #define for __builtin_ctzll incase it's not supported in the compiler.

I think @pingxie will merge it, although there are a bunch of test failures I am looking through. (Seem like some old issues, let me re-trigger them)

This is a breaking change IIRC, if there are existing mechanisms for log rotation we need to send a signal to the process to close and re-open the log. We...

@vax-r Ideally the todo list would be moved to github issues.