Skip to main content

· 5 min read
Hulk Lin

Apache Kvrocks 2.16.0 lands with many improvements and security fixes — extending the SET command with native compare-and-set, adding operator-friendly traffic control via CLIENT PAUSE/UNPAUSE, sharpening cluster slot management with CLUSTERX FLUSHSLOTS, and growing the probabilistic-data-structure surface with TDIGEST.TRIMMED_MEAN.

Redis 7.0+ command compatibility moves another step forward through BITPOS BYTE/BIT range units, CLIENT SETINFO, and the LATENCY family. Under the hood, RocksDB is upgraded to v11.1.1, and a broad round of security hardening (Lua CVE patches, replication tightening, RDB validation) ships alongside.

Highlights:

  • Conditional SET operations (IFEQ/IFNE/IFDEQ/IFDNE) — Atomic compare-and-set patterns are now supported natively in SET, enabling optimistic locking without a GET/SET round-trip.
  • CLIENT PAUSE / UNPAUSE — Operators can pause client traffic during failover windows without closing connections.
  • CLUSTERX FLUSHSLOTS — Drop all keys in one or more slot ranges in a single command — useful for slot rebalancing and tenant cleanup in cluster mode.
  • TDIGEST.TRIMMED_MEAN — Compute a trimmed mean over a t-digest sketch, completing the streaming quantile/aggregation toolkit.
  • BITPOS with BYTE / BIT range units — BITPOS now matches Redis 7.0+ semantics for byte-indexed and bit-indexed ranges.

· 3 min read
Hulk Lin

This release marks a significant step forward in Redis compatibility: the new redis-databases option closes the gap of SELECT command between Kvrocks and Redis. Applications that rely on Redis's multi-database model can now migrate to Kvrocks without any code changes.

Beyond that, this release expands TimeSeries and TDigest command coverage and tightens replication reliability — all while upgrading the C++ standard to C++20 and RocksDB to v10.10.1.

Highlights:

  • Support SELECT database — Kvrocks can now emulate Redis's multi-database model, unblocking migrations from Redis deployments that use SELECT to switch between logical databases
  • New TimeSeries commands — TS.MREVRANGE, TS.QUERYINDEX, and TS.ALTER round out the time-series API surface for analytics-heavy workloads
  • New TDigest commands — TDIGEST.RANK, TDIGEST.REVRANK, TDIGEST.BYRANK, and TDIGEST.BYREVRANK enable percentile and rank queries on streaming data

· 3 min read

The year 2023 is a small milestone for the community in open source, and it is also a new starting point. Kvrocks successfully graduated from the incubator to become an Apache top-level project in June, which is a great affirmation for the community's health and sustainability. At the same time, there are also some exciting progress in the project and community.

Before we start, we would like to thanks all the contributors and users who have been supporting Apache Kvrocks. We can't make it without you!

· 8 min read
Hulk Lin

Kvrocks is an open-source key-value database that is based on rocksdb and compatible with Redis protocol. Intention to decrease the cost of memory and increase the capability while compared to Redis. We would focus on how we use RocksDB features to improve the performance of the Redis on disk. Hopes this helps people who want to improve performance on RocksDB.

· 7 min read
Hulk Lin

Most developers should be familiar with bitmap, in addition to the storage implementation for the bloom filter, and many databases also provide bitmap type indexes. For memory storage, the bitmap can be regarded as the special type of sparse bit array, which would not cause the read-write amplification issue (means read/write bytes far more than the request). While Redis supports bit-related operations on string types, it is a big challenge for disk KV-based storage like Kvrocks. So this article mainly discusses "How to reduce disk read/write amplification on RocksDB".