All Versions
125
Latest Version
Avg Release Cycle
52 days
Latest Release
871 days ago
Changelog History
Page 12
Changelog History
Page 12
-
v2.0.0-rc6 Changes
April 19, 2018๐ Improvements
- ๐ Reduce lock contention in Worker
- โ Add metrics to the FuturePool ### ๐ Bug Fixes
- ๐ Fix misused metrics in Coprocessor
-
v2.0.0-rc.5 Changes
April 17, 2018๐ New Features
- ๐ Support compacting Regions in
tikv-ctl
- โ Add raw batch put/get/delete/scan API for TiKV service
- โ Add ImportKV service
- ๐ Support eval error in Coprocessor
- ๐ Support dynamic adjustment of RocksDB cache size by
tikv-ctl
- Collect number of rows scanned for each range in Coprocessor
- ๐ Support treating overflow as warning in Coprocessor
- ๐ Support learner in raftstore ### ๐ Improvements
- โฑ Increase snap GC timeout
- ๐ Support compacting Regions in
-
v2.0.0-rc.4 Changes
April 01, 2018๐ New Features
- Limit the memory usage during receiving snapshots, to avoid OOM in extreme conditions
- ๐ Support configuring the behavior of Coprocessor when it encounters warnings
- ๐ Support importing the data pattern in TiKV
- ๐ Support splitting Region in the middle ### ๐ Improvements
- ๐ Fix the issue that too many logs are output caused by leader missing when TiKV is isolated
- ๐ท Use crossbeam channel in worker
-
v2.0.0-rc.3 Changes
March 23, 2018๐ New Features
- ๐ Support Region Merge
- โ Add the Raw DeleteRange API
- โ Add the GetMetric API
- ๐ Support streaming in Coprocessor
- ๐ Support modifying RocksDB parameters online ### ๐ Improvements
- Inform PD immediately once the Raft snapshot process is completed, to speed up balancing
- โฌ๏ธ Reduce the I/O fluctuation caused by RocksDB sync files
- โก๏ธ Optimize the space reclaiming mechanism after deleting data
- ๐ Improve the data recovery tool
tikv-ctl
- ๐ Fix the issue that it is slow to make nodes down caused by snapshot
- Increase the raw_get/get/batch_get by 30% with ReadPool
- ๐ Support configuring the request timeout of Coprocessor
- Carry time information in Region heartbeats
- Limit the space usage of snapshot files to avoid consuming too much disk space
- Record and report the Regions that cannot elect a leader for a long time
- Speed up garbage cleaning when starting the server
- โก๏ธ Update the size information about the corresponding Region according to compaction events
- ๐ Limit the size of scan lock to avoid request timeout
- ๐ Use DeleteRange to speed up Region deletion
-
v2.0.0-rc.2 Changes
March 15, 2018๐ New Features
- Implement IngestSST API
tikv-ctl
now can send consistency-check requests to TiKV- ๐ Support dumping stats of RocksDB and malloc in
tikv-ctl
### ๐ Improvements - Reclaim disk space after data have been deleted
-
v2.0.0-rc.1 Changes
March 09, 2018๐ New Features
- ๐ง Protect important configuration which cannot be changed after initial configuration
- Check whether SSD is used when you start the cluster ### ๐ Improvements
- ๐ Fix the issue that gRPC call is not cancelled when PD leaders switch
- ๐ Optimize the read performance using ReadPool, and improve the performance by 30% for raw get
- ๐ Improve metrics and optimize the usage of metrics
-
v1.1.0-beta Changes
February 24, 2018๐ Improvements
- Traverse locks using offset + limit to avoid potential GC problems
- ๐ Support resolving locks in batches to improve GC speed
- ๐ Support GC concurrency to improve GC speed
- โก๏ธ Update the Region size using the RocksDB compaction listener for more accurate PD scheduling
- โ Delete the outdated data in batches using DeleteFilesInRanges, to make TiKV start faster
- ๐ง Configure the Raft snapshot max size to avoid the retained files taking up too much space
- ๐ Support more recovery operations in tikv-ctl
- โก๏ธ Optimize the ordered flow aggregation operation
-
v1.1.0-alpha Changes
January 19, 2018๐ New Features
- ๐ Support Raft learner
- ๐ Support TLS ### ๐ Improvements
- โก๏ธ Optimize Raft Snapshot and reduce the I/O overhead
- ๐ Optimize the RocksDB configuration to improve performance
- ๐ Optimize count (*) and query performance of unique index in Coprocessor
- Solve the reconnection issue between PD and TiKV
- โจ Enhance the features of the data recovery tool
tikv-ctl
- ๐ Support the Delete Range feature
- ๐ Support splitting according to table in Regions
- ๐ Support setting the I/O limit caused by snapshot
- ๐ Improve the flow control mechanism
-
v1.0.8 Changes
February 11, 2018๐ Improvements
- ๐ Use DeleteFilesInRanges to clear stale data and improve the TiKV starting speed
- ๐ Sync the metadata of the received Snapshot compulsorily to ensure its safety ### ๐ Bug Fixes
- ๐ Use Decimal in Coprocessor sum
-
v1.0.7 Changes
January 22, 2018๐ Improvements
- ๐ Support key-only option in Table Scan executor
- ๐ Support the remote mode in tikv-ctl
- ๐ Fix the loss of scheduling command from PD ### ๐ Bug Fixes
- ๐ Fix the format compatibility issue of tikv-ctl proto
- โ Add timeout in Push metric