LoginSign Up
← Back to all sparks
I

InfluxDB

ANALYTICS
Velocity5.0

Open-source time-series database for metrics and events

www.influxdata.com ↗

InfluxDB keeps patching its 3.11 and 3.12 lines, now with a catalog repair tool.

time-seriesdata-correctnesscatalogenterprisepatch releases
◆Current state
InfluxDB 3.12.0 made authentication and RBAC generally available in Enterprise and added distributed compaction in beta. Since then the team has shipped patch pairs for 3.12 and 3.11: lost-write and catalog-load fixes in mid-October, and on 10-29 a catalog deadlock fix, correct column names for WAL plugins, and 429 instead of 500 when a query runs out of memory. 3.12.2 also gets the offline catalog repair command.
◆Where it's heading
Most recent work is about data correctness and recovery, not new features. Two release lines get the same fixes, and an offline repair tool suggests catalog corruption has reached real users.
◆Prediction
More paired 3.12.x and 3.11.x patches are likely before any 3.13 feature release, since every release since 3.12.0 has been a fix.

◆Recent moves

  1. 11d ago

    InfluxDB 3.12.2 adds catalog repair, fixes deadlock

    3.12.2 adds the offline catalog repair command for catalogs broken by duplicate names or reused IDs, fixes a catalog deadlock, and returns 429 when a query runs out of memory. It continues the recovery-focused patches since 3.12.0.

    View source ↗
  2. 11d ago

    InfluxDB 3.11.8 backports catalog deadlock and 429 fixes

    The 3.11 line gets the same deadlock, WAL plugin column-name, and out-of-memory 429 fixes as 3.12.2. Users who haven't upgraded stay covered.

    View source ↗
  3. 29d ago

    InfluxDB 3.12.1 fixes lost writes and post-upgrade catalog failures

    The first 3.12 patch fixes a catalog that blocked startup after upgrading, acknowledged writes lost after a forced snapshot and restart, and overwrites resolving to the older value. A needed follow-up that shows the storage-correctness work continuing under the new features.

    View source ↗
  4. 29d ago

    InfluxDB 3.11.7 backports lost-write and catalog-load fixes

    The 3.11 line gets the same catalog-load and lost-write fixes, plus Enterprise fixes for duplicate node IDs and ingest nodes that wouldn't restart after an unclean stop. Operators who aren't ready for 3.12's one-way catalog upgrade still get the data-loss fixes.

    View source ↗
  5. 29d ago

    InfluxDB 3.12: GA auth and RBAC, distributed compaction beta

    3.12 is the first feature release in this run of patches: user authentication and RBAC reach GA in Enterprise, distributed compaction enters beta, and schema enforcement and faster bulk import arrive. It confirms the auth GA the patch lines had been preparing, but the one-way catalog upgrade means a backup is required before trying it.

    View source ↗
  6. 29d ago

    InfluxDB 3.11.5 adds age-based snapshots and upgrade metrics

    An age-based snapshot trigger stops low-traffic nodes from holding hours of unsnapshotted data, and Enterprise gains WAL, snapshot and stalled-upgrade metrics. Useful operational visibility for the Parquet engine migration, though the forced-snapshot path later needed a lost-write fix.

    View source ↗