ZanRedisDB v0.9.4

• speedup and fix snapshot recovery
• avoid leader transfer while applying the snapshot
• avoid transfer snapshot while not in sync only mode

ZanRedisDB v0.9.3

1. fix max watched index for the cluster meta info
2. precheck some write before propose to optimize
3. add memory radix index for test
4. add a config for slow limiter refuse cost
5. use v2 bitmap by default
6. use zap logger by default
7. adjust the keepalive for connection

ZanRedisDB v0.9.2

1. support lazy clean compacted TTL
2. zrem/srem optimized to avoid the raft
3. add the API to support trigger the compaction for the expired meta and other range
4. support level_compaction_dynamic_level_bytes for rocksdb
5. range del can be configured to ignore to sync to the remote cluster
6. fix placedriver leader election under network timeout
7. use hard-link for remote backup to reduce disk usage.

ZanRedisDB v0.9.1

1. add timeout for rsync transfer
2. try compact the expired meta if scan cost long
3. add a new balance strategy to reduce data migrate
4. use the magic code and check the magic code while start to avoid use old deleted namespace

ZanRedisDB v0.9.0

1. The rockredis support choose other key-value storage engines (currently the pebble/memory storage added for the test)
2. add getrange/set ex nx/setifeq/delifeq command
3. add stale read command to allow hot read from client
4. slow limiter support pre wait on queue
5. The cluster sync learner support enable/disable sync without a restart.
6. support send slow log to remote
7. merge some fix from etcd which will allow restarting on some corrupt data.
8. make sure the TTL cleaner not too busy.

ZanRedisDB v0.8.4

1. add metric for topn write and topn large collections
2. add slow logs for slow write and large collections
3. update the rocksdb to v6.4.6

ZanRedisDB v0.8.3

1. add more Prometheus metrics for latency and event
2. fix json.get & json.del for whole key
3. make slow limiter conf dynamic
4. optimize the collection clear while using compact TTL
5. fix ltrim with start>end or start > llen for empty list
6. fix spop response with no count
7. sclear/lclear/zclear now return the 1 if cleared.

ZanRedisDB v0.8.2

add slow limiter to refuse slow write in state machine while apply slowing

ZanRedisDB v0.8.1

1. optimize spop and some range command
2. use glog for logging
3. fix propose error while no leader
4. stats API do not return tables by default
5. handle empty proposal in remote cluster syncer

ZanRedisDB v0.8.0

1.support compacted TTL which will only clean expired keys during db compact to improve the performance for ttl (disabled by default)
2.use go mod
3.merge raft fix and optimize from etcd
4.refactor Redis API handler
5. use segmented bitmap for bit op to optimize the performance for big bitmap (disabled by default)
6.avoid receive raft logs if apply fall behand too much to reduce memory cost
7. optimize the cluster sync use pipeline and fix cluster synced term-index while restore from snapshot
8. add srandmember command
9.add a new Redis propose for the raft to reduce memory alloc and keep namespace in the raft proposal (disabled by default)