You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: site/en/release_notes.md
+133Lines changed: 133 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -8,6 +8,139 @@ title: Release Notes
8
8
9
9
Find out what’s new in Milvus! This page summarizes new features, improvements, known issues, and bug fixes in each release. You can find the release notes for each released version after v2.6.0 in this section. We suggest that you regularly visit this page to learn about updates.
10
10
11
+
## v2.6.6
12
+
13
+
Release date: November 21, 2025
14
+
15
+
| Milvus Version | Python SDK Version | Node.js SDK Version | Java SDK Version | Go SDK Version |
We are excited to announce the release of Milvus 2.6.6, featuring a range of powerful new capabilities, performance enhancements, and essential bug fixes. This update introduces important features such as Geospatial and Timestampz data type, Boost ranker for rescoring, etc. This release also has many crucial scalar filtering performance improvements. Several critical bugs have also been addressed to ensure greater stability and reliability. With this release, Milvus continues to provide a more robust and efficient experience for all users. Below are the key highlights of this release.
20
+
21
+
- Geospatial Data Type: Milvus introduces support for the `Geometry` data type, representing OGC-compliant geometric objects such as `POINT`, `LINESTRING`, and `POLYGON`. This type supports multiple spatial relationship operators (st_contains, st_intersects, st_within, st_dwithin, …) and provides an `RTREE` spatial index to accelerate spatial filtering and query execution. This enables efficient storage and querying of geospatial shapes for LBS, mapping, and other spatial workloads.
22
+
- Timestamptz Data Type: Milvus introduces the TIMESTAMPTZ data type, providing timezone awareness for all temporal data. This feature enables consistent data management across global deployments by allowing users to define a default time context using the timezone property on Databases and Collections. Crucially, the field fully supports expression-based filtering for time range queries, and retrieval operations (query and search) support a timezone parameter for instant, on-the-fly conversion of timestamps into the required local format upon output.
23
+
- Boost Ranker: Instead of relying solely on semantic similarity calculated based on vector distances, Boost Ranker allows Milvus to use the optional filtering condition within the function to find matches among search result candidates and boosts the scores of those matches by applying the specified weight, helping promote or demote the rankings of the matched entities in the final result.
24
+
- STL_SORT index now supports VARCHAR and TIMESTAMPTZ datatype.
25
+
- You may now enable dynamic field of an existing collection by altering it.
26
+
- Fixed cve-2025-63811.
27
+
28
+
### Features
29
+
30
+
- Added new config and enabled dynamic update configs ([#45363](https://github.com/milvus-io/milvus/pull/45363))
- Prevented panic when streaming coord shutdown but query coord still worked ([#45696](https://github.com/milvus-io/milvus/pull/45696))
81
+
- Set task init when worker didn’t have task ([#45676](https://github.com/milvus-io/milvus/pull/45676))
82
+
- Prevented deadlock in runcomponent when prepare failed ([#45647](https://github.com/milvus-io/milvus/pull/45647))
83
+
- Prevented panic when double closing channel of ack broadcast ([#45662](https://github.com/milvus-io/milvus/pull/45662))
84
+
- Corrected default value backfill during addfield ([#45644](https://github.com/milvus-io/milvus/pull/45644))
85
+
- Compacted the assignment history of channel to decrease the size of assignment recovery info ([#45607](https://github.com/milvus-io/milvus/pull/45607))
86
+
- Handled default values correctly during compaction for added fields ([#45619](https://github.com/milvus-io/milvus/pull/45619))
87
+
- Removed validatefieldname in dropindex ([#45462](https://github.com/milvus-io/milvus/pull/45462))
88
+
- Ignored compaction task when from segment was not healthy ([#45535](https://github.com/milvus-io/milvus/pull/45535))
89
+
- Set schema properties before broadcasting alter collection ([#45529](https://github.com/milvus-io/milvus/pull/45529))
90
+
- Stored database event if the key was invalid ([#45530](https://github.com/milvus-io/milvus/pull/45530))
91
+
- Fixed bulkimport bug for struct field ([#45536](https://github.com/milvus-io/milvus/pull/45536))
92
+
- Failed to get raw data for hybrid index ([#45408](https://github.com/milvus-io/milvus/pull/45408))
93
+
- Retained collection early to prevent it from being released before query completion ([#45415](https://github.com/milvus-io/milvus/pull/45415))
94
+
- Used the right resource key lock for ddl and used new ddl in transfer replica ([#45509](https://github.com/milvus-io/milvus/pull/45509))
95
+
- Fixed index compatibility after upgrade ([#45374](https://github.com/milvus-io/milvus/pull/45374))
96
+
- Fixed channel not available error and released collection blocking ([#45429](https://github.com/milvus-io/milvus/pull/45429))
97
+
- Removed collection meta when dropping partition ([#45497](https://github.com/milvus-io/milvus/pull/45497))
98
+
- Fixed target segment marked dropped for save stats result twice ([#45479](https://github.com/milvus-io/milvus/pull/45479))
99
+
- Wrongly updated timetick of collection info ([#45471](https://github.com/milvus-io/milvus/pull/45471))
100
+
- Added tzdata dependency to enable iana time zone id recognition ([#45495](https://github.com/milvus-io/milvus/pull/45495))
101
+
- Corrected field data offset calculation in rerank functions for bulk search ([#45482](https://github.com/milvus-io/milvus/pull/45482))
102
+
- Fixed filter geometry for growing with mmap ([#45465](https://github.com/milvus-io/milvus/pull/45465))
103
+
- Nextfieldid did not consider struct ([#45438](https://github.com/milvus-io/milvus/pull/45438))
104
+
- Group value was nil ([#45419](https://github.com/milvus-io/milvus/pull/45419))
105
+
- Provided accurate size estimation for sliced arrow arrays in compaction ([#45352](https://github.com/milvus-io/milvus/pull/45352))
106
+
- Fixed data race in replicate stream client ([#45347](https://github.com/milvus-io/milvus/pull/45347))
107
+
- Skipped building text index for newly added columns ([#45317](https://github.com/milvus-io/milvus/pull/45317))
108
+
- Accidentally ignored sealed segments in l0 compaction ([#45341](https://github.com/milvus-io/milvus/pull/45341))
109
+
- Moved finishload before text index creation to ensure raw data availability ([#45335](https://github.com/milvus-io/milvus/pull/45335))
110
+
- Did not use json_shredding for json path is null ([#45311](https://github.com/milvus-io/milvus/pull/45311))
111
+
- Cherry-picked fixes related to timestamptz ([#45321](https://github.com/milvus-io/milvus/pull/45321))
112
+
- Fixed load segment failure due to get disk usage error ([#45300](https://github.com/milvus-io/milvus/pull/45300))
113
+
- Supported json default value in compaction ([#45331](https://github.com/milvus-io/milvus/pull/45331))
114
+
- Computed the correct batch size for the geometry index of the growing segment ([#45261](https://github.com/milvus-io/milvus/pull/45261))
0 commit comments