Skip to content

Orchestrator v4.31.1

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 28 Jul 12:53
v4.31.1
80ac2f9
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Orchestrator v4.31.1 Release Notes

Orchestrator v4.31.1 is a corrective release following v4.31.0. Please use v4.31.1 as the authoritative release for this release cycle.

This release includes the reliability, observability, compatibility, and web UI improvements introduced since v4.30.2, together with the now-merged post-failover hook policy described below.

Important Correction

The v4.31.0 tag was created from an incorrect pre-merge commit. Its release page has been marked as superseded. v4.31.1 is tagged from master after PR #115 was merged and is the recommended version to install.

New Configuration Option

Make post-failover hook failures fail recovery

Failover is often more than promoting a new database primary. Deployments may also need to update service discovery, DNS, a load balancer, monitoring, or another external system. If that post-failover automation fails, the database can be healthy while the service is still not usable.

The new FailRecoveryOnFailedPostFailoverProcesses configuration option lets operators decide whether a non-zero post-failover process should cause the recovery to be marked unsuccessful.

  • The default is false, preserving existing behavior.
  • When enabled, failures from the relevant synchronous post-failover processes are treated as recovery failures.
  • Asynchronous hooks remain best-effort.
  • Topology changes are not rolled back when a post-failover process fails.

This allows each deployment to define explicitly whether its post-failover automation is part of the recovery contract.

See the recovery configuration documentation before enabling the option: docs/configuration-recovery.md.

Observability

Prometheus metrics now provide better visibility into Orchestrator itself, including health, monitored instance counts, discovery activity, recovery counts, and recovery duration.

These signals help operators distinguish a database health problem from a problem in the automation responsible for discovering topology or carrying out recovery.

Database Reliability

  • PostgreSQL graceful primary switchover support and functional validation were improved. The tests validate the database state directly and support round-trip switchover testing.
  • MariaDB DeadMaster detection and relay-log draining were hardened to avoid false positives during planned operations and to make replica repointing cleaner.
  • MySQL 9.7 LTS was added to the functional test matrix alongside MySQL 5.7, 8.0, 8.4, and 9.6.
  • Functional coverage continues across MariaDB, PostgreSQL 15/16/17, and Raft.

Web UI, Dependencies, and Build Improvements

  • Restored the web UI {{yield}} rendering behavior that caused HTTP 500 responses on affected pages.
  • Fixed cases where orchestrator -version produced empty output by ensuring the release version is embedded consistently during builds.
  • Upgraded bundled jQuery to 3.7.1 and jQuery UI to 1.13.3.
  • Improved CI reliability, linting, runtime image caching, and package/build validation.

Upgrade Guidance

Review the new post-failover policy before enabling it. Existing deployments retain the previous behavior unless FailRecoveryOnFailedPostFailoverProcesses is set to true.

Full comparison with the previous release line: v4.30.2...v4.31.1

Assets 16
Loading