Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

In some states `raft-rs` ignores the ReadIndex request.
Check it preliminary, don't wait for the timeout.

See for details:
- <https://github.com/tikv/raft-rs/blob/v0.6.0/src/raft.rs#L2058>
- <https://github.com/tikv/raft-rs/blob/v0.6.0/src/raft.rs#L2323>

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

32MiB is the minimum allowed memtx_memory size.

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

No need to DDoS a poor raft leader, especially when it's already under
high load.

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

The failure of `raw_node.propose_conf_change` isn't easy to handle.
For now we only know that it shouldn't short-circuit the
`raft_conf_change_loop`. Fuse it with a `fiber.sleep()`.

Current pause value of 1 second is chosen arbitrarily with no actual
reason.

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

Join requests should be made without timeout restrictions. Otherways
it's impossible to tell retried request from instance_id collision.
Retried requests never succeed and return the "already joined" error.

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

There's no phenomenon like replicaset mismatch nowadays. The cases this
test actually reproduces are already covered by test_override.

It makes no sence to perform a wait, and `docs/clustering.md` already
states that it's skipped. Anyway, the result is unused.

The case is wierd, but it could lead to box error and panic.

1. Merge duplicating tests
2. Provide detailed comments

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

Part of https://git.picodata.io/picodata/picodata/picodata/-/issues/98

Yaroslav Dynnikov authored 2 years ago and Georgy Moshkin committed 2 years ago

Part of https://git.picodata.io/picodata/picodata/picodata/-/issues/98

Initially, JoinResponse was supplied with every single peer in cluster.
If the cluster is big enough, it results in the "max tuple size limit
reached" error. In fact, only voters are needed for the raft to
operate normally.

Since a recent commit, message destination (`to` field) is taken into
account. Make it correct to avoid skipping.

Joint state latch was mistakenly dropped, and conf_change_loop couldn't
receive a notification.

Also assert it's not overridden by mistake.