matrixcube
matrixcube copied to clipboard
Fundamental Building Block for Elastic Storage With Strong Consistency and Reliability
In the current implementation, shard metadata is not stored within shard's range, e.g. consider a KV data storage with a fixed range, there is currently no mechanism to ensure its...
Review all fields in meta.RaftMessage to see is there any redundant fields that can be removed to shrink serialized size.
Some changes have been made in #297, still some work left to stop the replica from referencing the store instance.
Three kinds of configuration could be possible for different kinds of cases: - Default strong consistency - Two replica one of which is raft learner - No RSM, only single...
both etcd and pebble preallocate disk spaces when writing logs. preallocated disk space need to be configurable for tests.
Support custom skip apply raft log
Currently different slice groups use a split rule, this is clearly inappropriate. We need isolation of splitting rules for different shard groups
Support opentracing for future bug tracking, tuning, monitoring, etc.
Assuming gossip would be the approach to fully decentralize the PD service, we'd like to know - 1. what stats need to be collected and shared across the cluster. 2....
We need more precise retry error types to minimize the latency of requests