MariaDB 13.0 Reaches Stable With a Different Upgrade Story From LTS
MariaDB 13.0.2 became the stable rolling release in September 2026, adding SQL, optimizer, observability and storage improvements while remaining a shorter-lived branch than MariaDB's LTS releases.
MariaDB 13.0 is now stable, but it is a stable rolling release rather than an LTS branch, so the release is aimed at teams that want its newer SQL and operational capabilities without assuming a long maintenance window. MariaDB 13.0.2 reached Stable status in September 2026 after moving through Preview and Release Candidate stages.
That distinction matters more than the version number.
Stable does not mean LTS
MariaDB currently uses different release tracks.
The Foundation's release table shows:
- MariaDB 13.0.2 — Stable
- MariaDB 13.1.1 — Release Candidate
- MariaDB 12.3.3 — Stable
- MariaDB 11.8 — latest LTS series
The important concept is:
stable = maturity of this release
LTS = length and pace of supported maintenance
A stable rolling release can be ready for production while still requiring more frequent upgrading than an LTS branch.
That is exactly how MariaDB describes 13.0.
What changed between RC and Stable
MariaDB's own release note is unusually useful because it emphasizes what happened after the headline features were introduced.
The project says the 13.0 branch received:
- 604 additional commits after RC;
- 1,701 changed files;
- 335 MDEVs addressed;
- and 23,748 additional test lines.
Those are project-reported development metrics, not an independent measure of release quality.
The useful point is qualitative.
The Stable milestone represents another stage of engineering work beyond the initial feature preview.
The release is about SQL and operational visibility
MariaDB 13.0's feature set includes several changes aimed at development and database operations.
The release materials highlight areas such as:
- UPDATE ... RETURNING;
- REF CURSOR and RECORD support for Oracle-compatible procedural SQL;
- optimizer hints and metadata visibility;
- InnoDB log archiving;
- replication and binary-logging improvements;
- and additional observability work.
That is a different story from a single breakthrough database feature.
The release is broad.
Oracle compatibility is a practical migration signal
One of the more visible changes is expanded compatibility with Oracle-style procedural SQL.
MariaDB 13 adds native support for REF CURSOR and RECORD types in routine parameters and return values.
That matters because compatibility is often less about syntax than migration friction.
A database team deciding whether to move a workload has to measure:
application SQL
+ stored procedures
+ operational tooling
+ drivers
+ replication
+ monitoring
Adding familiar constructs can reduce one part of the migration burden.
It does not prove that an arbitrary Oracle workload can migrate unchanged.
InnoDB log archiving shows the operational side
Not every important database feature is visible to application developers.
MariaDB's 13.0 work also expands operational capabilities, including InnoDB log archiving.
That feature matters in backup and recovery workflows because transaction-log handling can affect how administrators capture a consistent stream of changes around backup operations.
The broader direction is visible:
13.0 is not only trying to add SQL syntax. It is trying to make the server more observable and operable.
The "stable rolling release" model changes upgrade decisions
An LTS branch encourages long periods between major feature changes.
A rolling branch offers newer capabilities more quickly but expects a more active upgrade process.
That creates two different operational strategies:
LTS strategy
|
slower changes
|
longer validation windows
|
less frequent upgrades
Rolling strategy
|
newer features
|
shorter lifecycle
|
more active upgrade cadence
Neither is inherently better.
The right choice depends on how much an organization values feature velocity relative to maintenance stability.
MariaDB 13.0 sits inside a broader database ecosystem
The Observatory recently covered Google Cloud's native BM25 support in PostgreSQL.
That article focuses on a specific retrieval workload being pulled closer to the database layer.
MariaDB 13.0 represents a broader database-engine evolution: SQL compatibility, observability, storage and optimizer improvements.
The two observations point to the same pattern:
databases are becoming more specialized application platforms rather than passive persistence layers.
They are increasingly expected to handle search, analytics, operational telemetry and application-specific data behavior.
Why version maturity matters more than the feature list
Database release notes often read like a checklist.
But the operational question is:
Is the release mature enough for this workload?
A stable tag provides some evidence because the project has passed its RC stage.
It does not substitute for application testing.
A production migration still needs:
- query regression tests;
- replication tests;
- driver compatibility;
- backup and restore validation;
- schema migration testing;
- performance baselines;
- and rollback planning.
MariaDB itself explicitly recommends testing 13.0 against actual workloads before upgrading.
Why this is meaningfully new for Digital Observatory
The Observatory did not have coverage of MariaDB 13.0, its rolling-release maturity model or the specific gap between Stable and LTS release tracks.
That makes this a useful addition because release cadence itself is an infrastructure signal.
Database teams do not only choose features.
They choose an upgrade policy.
What to watch next
- Whether MariaDB 13.0 receives real-world adoption despite its shorter support horizon.
- How the Oracle compatibility additions behave on large procedural workloads.
- Whether the rolling-release model becomes easier for managed database providers to operationalize.
- How MariaDB 13.1 moves from RC to Stable.
- Whether workload-aware testing becomes more important as databases ship larger quarterly feature sets.
Limitations
The commit and test figures in this article are project-reported measurements. They show the amount of stabilization work, not a proven quality score.
Stable status also does not mean every application will upgrade cleanly. Compatibility, workload behavior and performance still need to be tested in the target environment.
Sources and further reading
- MariaDB Foundation — MariaDB 13.0 is now Stable
- MariaDB Foundation — all releases
- MariaDB announce mailing list — 13.0.2
- InfoQ — MariaDB 13 release
- Linuxiac — software news
Related Digital Observatory reads: Google Cloud native BM25 in PostgreSQL, PyTorch 2.14 distributed fault tolerance, and Rust 1.98.1 release-process signal.
Evidence
Sources & further reading
Primary sources, official disclosures, and external research used to ground this report.
- MariaDB Foundation — MariaDB 13.0 is now Stablemariadb.org
Primary September 17, 2026 release announcement describing the stable milestone, rolling-release status and additional stabilization work.
- MariaDB Foundation — all releasesmariadb.org
Primary release table showing MariaDB 13.0.2 as Stable and MariaDB 13.1.1 as the next release candidate.
- MariaDB announce mailing list — 13.0.2mail-archive.com
Public release announcement for MariaDB 13.0.2 and 13.1.1 with the high-level feature list.
- InfoQ — MariaDB 13 releaseinfoq.com
Independent September 18, 2026 coverage of MariaDB 13.0 GA and its SQL/Oracle-compatibility improvements.
- Linuxiac — MariaDB 13.0linuxiac.com
Independent September 18, 2026 software-news index corroborating the stable release and highlighted feature changes.
Keep Exploring
Related observations.
Kubeflow's CNCF Graduation Makes Kubernetes a More Credible AI Platform Boundary
Kubeflow's graduation is not a new benchmark or model release; it is an ecosystem-maturity signal that changes how platform teams can evaluate Kubernetes-native AI infrastructure.
SQL Joins, GROUP BY, and Aggregation: The Relational Thinking Students Need
SQL is easiest to understand when you think in relations and transformations rather than translating every query into an imaginary nested loop.
SQL Injection and Safe Database Queries: Why Parameters Beat String Concatenation
SQL injection happens when untrusted input changes the structure of a database command. Parameterized queries keep data in the data channel.