Jiaming Cao comments

Results 9 comments of


                                            Jiaming Cao

Possible split-brain & loss of committed write with loss of un-fsynced data

My guess: [`MustSync` ](https://github.com/etcd-io/raft/blob/4abd9e927c6d5db930dfdb80237ac584449aeec7/rawnode.go#L193-L200)does not check commit index, and the updates of commit index is not fsync-ed. This violates the [usage](https://github.com/etcd-io/raft#usage) of raft: > Write Entries, HardState and Snapshot to...

Implement a self-contained raft example

Can I take this over?

Exploratory test for freepages allocator

Hi, can I try to work on this?

Exploratory test for freepages allocator

@ptabor Sorry if I misunderstand something. How to performs a RW transaction when there is open RO transaction? Or do you mean performs a RW transaction only after all RO...

Release in-memory log space based on entries size

I am taking a look at this issue. However, from what I've learned, I don't think there is a way to get the length of entire backing array of a...

Release in-memory log space based on entries size

I didn't see https://github.com/etcd-io/raft/issues/69. I think this issue would be easy to solve if we get https://github.com/etcd-io/raft/issues/69 done.

Allow recovering from some assertions

I am trying to solve the commit index regression problem first. I am now writing an interaction test to reproduce the problem.

Allow reuse of slices exposed through Ready

@tbg @ahrtr Please see #51

Allow reuse of slices exposed through Ready

@pavelkalinnikov Thank you for your comments! I've resolved them. > we still reallocate and copy the slice. The trick is at when we put back the slice: ``` rn.raft.msgsAfterAppendPool.Put(rn.raft.msgsAfterAppend[:0]) ```...