NATS Files

Changelog

Refer to the 2.12 Upgrade Guide for backwards compatibility notes with 2.11.x.

Go Version

• 1.25.3 (#7428)

Dependencies

• github.com/google/go-tpm v0.9.6 (#7374)
• github.com/nats-io/nats.go v1.46.0 (#7374)
• golang.org/x/crypto v0.43.0 (#7423)
• golang.org/x/sys v0.37.0 (#7423)
• golang.org/x/time v0.14.0 (#7423)

Added

General

• The NewServerFromConfig function has been added to the embedded server API for parsing an external configuration file when ConfigFile is supplied (#7364)
• WebSocket leafnode connections can now use a HTTP proxy by specifying proxy settings in the leafnode remotes (#7242)
• The write_deadline option in the cluster, leafnode and gateway config blocks allows configuring write deadlines on a finger-grained basis (#7405)

Improved

Monitoring

• The jsz monitoring endpoint can now report leader counts (#7429)

JetStream

• The store lock is no longer held while searching for TTL expiry tasks, improving performance (#7344)
• Removing a message from the TTL state is now faster (#7344)
• The filestore no longer performs heap allocations for hash checks (#7345)
• The filestore now reuses pooled write cache allocations more efficiently (#7346)
• Meta snapshot performance for a very large number of assets has been improved after a regression in v2.11.9 (#7350)
• Sequence-from-timestamp lookups, such as those using opt_start_time on consumers or start_time on message get requests, now use a binary search for improved lookup performance (#7357)
• Scheduled messages are now deactivated properly when followed up with another message on the same subject without a schedule (#7366)
• Meta snapshots are no longer taken on every stream removal (#7373)
• Filestore cache lookups can now return early for empty blocks (#7381)
• Filestore meta files are now written using temporary staging, avoiding accidental truncation on crashes (#7388)
• Atomic batch publish now supports deduplication with the Nats-Msg-Id header (#7391)
• Cache-specific log lines in the filestore have been improved (#7396)
• Log lines and errors related to offline/unsupported assets are now clearer (#7416, [#7425])

Fixed

General

• Fixed the exit code when receiving a SIGTERM signal immediately after startup (#7367)
• The $SYS.REQ.CLAIMS.DELETE endpoint now correctly strips headers, like the $SYS.REQ.CLAIMS.UPDATE endpoint already does (#7413)
• Log lines generated when reloading trusted proxies now report the correct keys (#7427)

JetStream

• Batch check now uses the correct subject instead of the last subject that does the commit (#7342)
• Removed messages with a per-message TTL are now removed from the TTL state immediately (#7344)
• Fixed a bug where TTL state was recovered on startup with subject delete markers enabled, that message expiry would not start as expected (#7344)
• Expiring messages from the filestore no longer leaks timers and expires at the correct time (#7344)
• Fixed a bug where scheduled messages would not trigger correctly after recovery (#7347)
• Deleting a non-existent sequence on a stream no longer results in a cluster reset and leadership election (#7348)
• Subject tree intersection now correctly handles overlapping literals and partial wildcards, i.e. stream.A and stream.*.A, fixing some consumer or message get filters (#7349)
• A data race when checking all JetStream limits has been fixed (#7356)
• Raft will no longer trigger a reset of the clustered state due to a stream snapshot timeout (#7293)
• Atomic batches now reject unsupported commits (#7368)
• Race conditions and potential panics fixed in the handling of some JetStream API handlers (#7380)
• The filestore no longer loses tombstones when using secure erase (#7384)
• The filestore no longer loses the last sequence when recovering blocks containing only tombstones (#7384)
• The filestore now correctly cleans up empty blocks when selecting the next first block (#7384)
• The filestore now correctly obeys sync_always for writing TTL and scheduling state files (#7385)
• The filestore will now correctly expire a cache when the wrong sequence is found (#7396)
• Fixed a data race on a wait group when mirroring streams (#7395)
• Skipped message sequences are now checked for ordering before apply, fixing a potential stream desync on catchups (#7400)
• Skipped message sequences now correctly detect gaps from erased message slots, fixing potential cache issues, slow reads and issues with catchups (#7399, [#7401])
• Raft groups now report peer activity more consistently, fixing some cases where asset info and monitoring endpoints may report misleading values after leader changes (#7402)
• Raft groups will no longer permit truncations from unexpected catchup entries if the catchup is completed (#7424)
• The filestore will now correctly release locks when erasing messages returns an error (#7431)

Complete Changes

https://github.com/nats-io/nats-server/compare/v2.12.0...v2.12.1