What dragonflydb/dragonfly shipped

Dragonfly v2.1 is out. Starting with this version the project moves from BUSL-1.1 to the Dragonfly Source-Available License. Also this week: CLIENT KILL TYPE/USER/MAXAGE/SKIPME filters, search errors reported in FT.AGGREGATE and FT.HYBRID, HNSW graphs rebuilt on FT.ALTER, and:: a

Dragonfly v2.1 shipped this week. From this version the project moves from BUSL-1.1 to the Dragonfly Source-Available License, a change in how teams can use it in production. The week also brought server fixes and additions: CLIENT KILL now accepts TYPE, USER, MAXAGE and SKIPME filters, including the calls Sentinel sends when switching master; FT.AGGREGATE and FT.HYBRID now report shard errors the way FT.SEARCH does; FT.ALTER rebuilds HNSW graphs and retires dropped ones; EXPIREAT, PEXPIREAT and GETEX clamp over-cap deadlines consistently; and with no bind flags Dragonfly listens on:: by default on dual-stack hosts.

This week: cloud snapshot init failures no longer exit the server during SAVE or DFLY LOAD, ACL files fail closed and save atomically, TIME returns bulk strings, XINFO and JSON.DEBUG no longer hang scripts, and search index definitions survive reloads.

This week's changes focus on not losing data or availability by accident. A failure to initialize cloud snapshot storage during SAVE, BGSAVE or DFLY LOAD no longer calls exit, so a transient credential problem cannot kill a running instance. ACL file handling now fails closed and ACL SAVE writes atomically, so a concurrent load cannot read an empty file and wipe users. TIME replies with bulk strings as documented, fixing strict typed clients. XINFO and JSON.DEBUG are keyed commands, so scripts no longer hang waiting on their own shard. Search index definitions with spaces or wide schemas survive reloads and full sync, and a journaled FT.SYNUPDATE is kept on replicas. Cluster slot queries are cheaper, and post-migration slot flushing is throttled to keep foreground commands responsive.

This week: search SORTBY paging fixed, multiple attributes per JSON path, ACL categories for CLIENT/MEMORY, CONFIG SET secrets hidden from logs, SADD counting fixed, and AOF replay at startup.

This week's changes to Dragonflydb include several correctness and security fixes. Search SORTBY with LIMIT now orders ties by key and ranks missing values last, so paging no longer loses or duplicates documents. JSON indexes accept several attributes on one JSON path, and expired hash fields are removed from search indices. On the security side, CLIENT and MEMORY admin subcommands have their own ACL categories, RM respects key patterns, and CONFIG SET values no longer appear in logs. SADD now counts members correctly across integer-to-string set conversion. Stream commands other than XADD reject * as an explicit ID. AOF files are replayed at startup with a basic end-to-end test.

This week: search aliases on FT.ALIASADD/UPDATE/DEL, the event-driven RESP loop is now default, PFMERGE writes only on its own shard, expired keys report missing instead of WRONGTYPE, and a lagging replica flow can no longer starve its thread.

This week in Dragonfly: search indexes can now be managed through aliases with FT.ALIASADD, FT.ALIASUPDATE and FT.ALIASDEL. The event-driven RESP connection loop became the default. PFMERGE now writes its destination only on its own shard, fixing inflated DBSIZE, duplicate SCAN results and key count differences between master and replica. Expired keys are reported as missing instead of WRONGTYPE, FT.ALTER keeps synonym groups, and BM25 scores no longer change with the shard count. A replica flow that falls behind now yields so other fibers on its thread keep running, tiering fixes segment queue corruption, and V2 pipeline counters are exported through metrics.

V2.0.2 is out: fixes a crash when DFLYMIGRATE hits a non-cluster server and a crash on glob patterns with many * wildcards in KEYS and SCAN-family MATCH. This week also made deeply nested search queries iterative, and added an AOF segment writer with a per-shard streamer.

This week Dragonfly shipped v2.0.2, a patch release that fixes a crash when DFLYMIGRATE is sent to a server not in cluster mode and a crash on glob patterns with many * wildcards in KEYS and SCAN/HSCAN/SSCAN/ZSCAN MATCH patterns. Alongside the release: deeply nested search queries are now processed iteratively so they no longer overflow fiber stacks, search indices left unbuilt by a failed load are dropped instead of being reported as fully indexed, and snapshots under tiering wait for delayed entries taken over by writers before the final flush. Work on the new AOF on-disk format, block builder and segment writer continued.

New in Dragonfly: 1.88-3.36x faster double formatting with dragonbox, automatic OAHTable shrinking, real FT.INFO progress reporting, Valkey hash-expiry RDB support, zero-copy MGET, and punctuation in search queries.

This week's Dragonfly updates focus on performance and reliability. Double-to-string formatting is now 1.88-3.36x faster using dragonbox. OAHTable automatically shrinks after removals to reclaim capacity. Search queries gain real progress reporting and support punctuation like @email:*example.com*. MGET operations now borrow large values zero-copy. Replication becomes more robust by terminating unsupported RDB formats instead of retrying indefinitely. These improvements enhance both user experience and operational stability.

This week: cluster command batching reduces socket writes, SCRIPT FLUSH deadlock fix, FT.PROFILE error reporting, glob pattern overflow prevention, and tiering disk accounting improvements.

Weekly update: CLUSTER SLOTS/SHARDS now batch replies for massive write reduction on large configurations. SCRIPT FLUSH deadlock eliminated through non-blocking interpreter pool. Search profiling now reports shard errors correctly. Pattern matching prevents fiber stack overflow. Tiering accurately tracks large disk allocations. Plus fixes to empty field validation and replication test reliability.

Dragonfly v2.0.0 is here. Three years in the making. Expanded Redis/Valkey compatibility, optimized SCAN performance, better hash expiration accuracy, and critical reliability fixes.

We're excited to announce Dragonfly v2.0.0—a major milestone representing three years of development toward production maturity. This release includes optimized SCAN command performance through shared result storage, improved hash field expiration accuracy for deterministic replay, expanded Valkey TCL test compatibility covering 248 additional tests, and critical reliability fixes including proper CLIENT PAUSE cleanup and script loading contention reduction.