ktoso / akka-raft

A toy project implementing RAFT on top of Akka Cluster (not prod ready)

http://blog.project13.pl

ktoso/akka-raft Issues

Hacktoberfest
Closed 8 years ago2
Most of the tests are failing
Updated 8 years ago2
Any exception can cause two leaders to become elected in the same Term
Closed 8 years ago9
Leader should cancel SendHeartbeat timer when it steps down
Updated 8 years ago1
Perfomance issue: should overwrite log entries, even if AppendEntries is empty
Updated 9 years ago
Optimisation: Candidate can possibly retry on request vote if request delivered on lower currentTerm
Updated 9 years ago1
LogMatch invariant violated
Updated 9 years ago2
Would get better performance if akka-raft used udp
Updated 9 years ago4
votedFor Map grows without bound
Closed 9 years ago2
Add tests for merged fixup
Updated 9 years ago
Should only delete entries if they conflict with a new one
Closed 9 years ago11
Follower should be able to vote more than once in a term
Updated 9 years ago5
indexOnMajority is wrong
Updated 9 years ago3
Servers should check log consistency before granting vote
Updated 9 years ago1
Leader should contact majority before processing read-only client request
Updated 9 years ago
Client should attach serial numbers to ClientMessages
Updated 9 years ago
Leader should commit no-op entry upon election
Updated 9 years ago
Nodes should not forget who they voted for
Updated 9 years ago11
Log consistency check is incorrect
Updated 9 years ago1
RPC response not always sent
Updated 9 years ago
Candidate should have election restart timeout
Updated 9 years ago2
Follower does not check consistency of AppendEntries
Updated 9 years ago1
Follower does not commit entries on receiving heartbeat
Updated 9 years ago1
Election timeout reset twice on Follower->Candidate transition
Updated 9 years ago
DeclineCandidate from Candidate includes wrong term
Updated 9 years ago1
Incorrect next index calculation for log
Updated 9 years ago
Inconsistent use of index numbers
Updated 9 years ago1
termAt can return wrong term for first item
Updated 9 years ago1
LeaderTest starts log at wrong index
Updated 9 years ago
Log compaction can cause append to take too many entries
Updated 9 years ago
Empty batch is sent with term 1/index 1
Updated 9 years ago1
Follower does not reset election deadline after voting for a candidate
Updated 9 years ago1
Not all states handle all protocol messages
Updated 9 years ago
Candidate does not distinguish between votes from same follower
Updated 9 years ago2
Candidate should check term in VoteCandidate messages
Updated 9 years ago1
Members should not respond with Vote if sender is not in Configuration
Updated 9 years ago5
Delayed messages to self can cause leader to be elected without quorum
Updated 9 years ago1
All nodes should transition to Follower if term > currentTerm
Updated 9 years ago1
Shouldn't necessarily initialize matchIndex to -1
Updated 9 years ago2
Crash: "Unable to find log entry at index 2"
Updated 9 years ago1
Should reject AppendEntries if term < currentTerm
Closed 9 years ago
Follower should not vote if Candidate has out-of-date log
Closed 9 years ago1
Two Leaders Elected in the Same Term
Updated 9 years ago12
Raft based auto-downing for the cluster
Updated 10 years ago
Investigate ordering issue while replay
Closed 10 years ago
Simple clustered code example?
Closed 10 years ago1
Simple example continually logging: Consensus for persisted index: 1. (Comitted index: 1)...
Updated 10 years ago3
Update: 2 actors (cluster / raft) make it possible to leak raft "self", where clusterSelf needed, add checks?
Updated 10 years ago
Make raft actor names configurable
Closed 10 years ago1
Improve auto discovery and cluster resizes in clustered mode
Updated 10 years ago1