Fullstop000
Fullstop000
Currently, we use a simple file to store all log messages: https://github.com/Fullstop000/wickdb/blob/15c37ccff1c669a0f51831f4856afd374b631309/src/logger.rs#L54-L58 This could be a bad design as the log messages growing. Adding a file rotation by implementing a...
Add benchmark for components below : - [ ] LRU - [x] Arena - [x] Skiplist - [ ] utils/hash - [ ] Record - [ ] Table - [...
https://github.com/pingcap/raft-rs/issues/136 ## New features * Add some new fields in `Message` * New concepts `Group` and `Delegate` in the cluster topology * ~~A message now could be sent to a...
In `Raft::new`, we always create a `ProgressSet` with setting `next_idx` 1 in each `Progress`: https://github.com/tikv/raft-rs/blob/f7f7d0121132ef6406b59fc774778285cd1477d6/src/raft.rs#L251-L255 In the situation where recovering the raft node, I think maybe we can just use...
https://github.com/tikv/raft-rs/blob/f7f7d0121132ef6406b59fc774778285cd1477d6/src/raft.rs#L565-L566 `try_batch` doesn't check the entries size after batched which might cause the overflow. Do we need to worry about this? @hicqu @BusyJay
Part of https://github.com/tikv/raft-rs/issues/109 This PR tries to add a benchmark for the real raft node cluster. The communication between nodes is supported by `mspc::channel` which brings some overhead. @Hoverbear @hicqu...
It seems that `benches` are unable to use `harness` based on the current project structure. How about moving the entire `benches` into `harness`? @nrc @Hoverbear Now: ``` benches harness src...
> 如果primary lock 已被write所替代，也就是说该事务已被提交，事务需要roll forword,也就是对所有涉及到的、未完成提交的数据，用write记录替代标准的锁standard lock。 I think it should be `stranded lock` (remaining lock) but not `standard lock` referring to the last second paragraph in Chapter 2.2 of https://www.usenix.org/legacy/event/osdi10/tech/full_papers/Peng.pdf...
Relate to https://github.com/tikv/raft-rs/pull/249 Signed-off-by: Fullstop000