Skip to content

Release notes v2.3.4

Release highlights

  • Version name: ClusterControl 2.3.4
  • State: Latest
  • First release date: October 7th, 2025
  • Maintenance status: Active
  • Total maintenance builds: 7
  • Last release date: N/A
  • Restore external and partial backups operations for MySQL and PostgreSQL.
  • PostgreSQL custom data directory - Use any custom data directory location for your database nodes.
  • PostgreSQL local handling improvement for new deployment and major upgrades.
  • Kubernetes integration is now GA. See User Guide → Integration → Kubernetes.
  • Drop support for new cluster deployments for MariaDB-related clusters 10.4 and older.
  • "Enable TimescaleDB" as an PostgreSQL action/job has been deprecated.
  • PostgreSQL 18
    • Major upgrade from PostgreSQL 17 to 18 is not currently supported.
    • Import existing PostgreSQL 18 is not currently supported.
    • Only for streaming replication.
  • Early preview release of CMON controller pool only supports multiple controller pools co-located on a same host.
  1. Follow standard upgrade procedures as outlined in the documentation.

Maintenance Release: December 22nd, 2025

  • Build:

    • clustercontrol-mcc-2.3.4-693

  • GUI:

    • Address an issue (404) when opening the Kubernetes Settings sub pages (CLUS-6871)

Maintenance Release: December 17th, 2025

  • Build:

    • clustercontrol-controller-2.3.4-17322
    • clustercontrol-mcc-2.3.4-689
    • clustercontrol2-2.3.4-2187

  • Controller:

    • Address an issue with cmon exporter appending authentication entries to the pg_hba.conf file. (CLUS-6278)
    • Configure initdb to generate pg_hba.conf with peer authentication for local connections and md5 for host connections, replacing the default trust method.
    • CMON now prefers local socket for pg_basebackup when pg_hba.conf allows
    • Falls back to TCP/IP using the cmon_replication user and PGPASSWORD when local socket access is not allowed
    • Added support for positional insertion of pg_hba.conf entries instead of always appending
      • Added detection of local trust rules
      • CMON now inserts user-specific entries before any local trust rules
    • Improved default security posture by replacing global trust rules:
      • peer is now used for local (Unix socket) connections
      • md5 is enforced for host-based connections
    • Address an issue with expired MSSQL certificates and backup restore. Renew certificates with s9s CLI (CLUS-6631): s9s cluster --renew-cert --cluster-id=4
    • Address an improvement to update Prometheus version used to 2.53.5 LTS (CLUS-6829)
    • Address an issue with Valkey 7 deployments on Oracle 8 (CLUS-6818)
    • Address an issue with adding a new Valkey node (CLUS-6744)
    • Address issues creating a replica cluster with PostgreSQL (CLUS-6735)
    • Address improvements to hardened TLS configuration for cmon and cmnd (CLUS-6663, CLUS-6662)
      • Removed weak CBC and Camellia-based cipher suites from cmon and cmnd in line with current recommendations from https://ciphersuite.info.
      • New installations now enforce TLS 1.3 by default, while existing installations keep TLS 1.2 enabled for backward compatibility.
      • Added a new allow_tls12=true|false option in CMON’s configuration:
        • false: only TLS 1.3+ connections are allowed.
        • true: both TLS 1.2 and TLS 1.3+ are allowed.
      • The allow_tls12 setting is propagated to cmnd (cmnd.conf) to ensure consistent TLS behavior.
    • Address an issue with restoring external backup because primary could not be determined (CLUS-6590)
    • Address improvements with potential race conditions with frequent configuration changes and remove cluster operations (CLUS-6852)

  • GUI:

    • Address an issue where the bi-directional option was incorrectly shown for create replica cluster with PostgreSQL streaming replication cluster (CLUS-6796)
    • Address an issue with Query Monitoring/Agents page not refreshing when clicking on a different cluster (CLUS-6780)
    • Address an issue when editing backup schedules where storage host was not correct with MongoDB (CLUS-6757)
    • Address an improvement to truncate DB user privileges with there many PostgreSQL users listed (CLUS-6578)
    • Address an issue with a too narrow custom data directory input field with the PostgreSQL wizard (CLUS-6828)

Maintenance Release: December 2nd, 2025

  • Build:

    • clustercontrol-controller-2.3.4-17176
    • clustercontrol2-2.3.4-2180

  • Controller:

    • PostgreSQL locale handling improvements (CLUS-6693).

      • Fixed locale settings for PostgreSQL major upgrade jobs
      • No longer use LC_ALL=C for initdb commands.
      • Removed the requirement to explicitly set lc_collate, lc_ctype, and related locale settings to C. These values will now be automatically derived from server_encoding.
      • During upgrades, the system now reuses the previous node’s locale settings. Including lc_collate, lc_ctype, lc_messages, lc_monetary, lc_numeric, and lc_time. Ensuring that any custom locale configurations are preserved.

    • Added support for TimescaleDB 17 and 18 on the following platforms:

      • TimescaleDB 17 supported on:
        • Ubuntu 22.04 (Jammy),
        • Ubuntu 24.04 (Noble),
        • Debian 11 (Bullseye),
        • Debian 12 (Bookworm),
        • CentOS/RHEL 9.
      • TimescaleDB 18 supported on:
        • Ubuntu 22.04 (Jammy),
        • Ubuntu 24.04 (Noble),
        • Debian 12 (Bookworm),
        • CentOS/RHEL 9.
      • TimescaleDB 18 on Rocky Linux 9 is currently supported with limitations due to a Linux distribution issue.

    • Address a deployment issue affecting Valkey 8 and 9 on Rocky Linux 9. (CLUS-6748)

    • Address an issue when promoting a PostgreSQL replica to a primary (CCX)(CLUS-6573)

  • GUI:

    • Address an issue with PITR label with backup tool tip (CLUS-6577)
    • Address an issue with inconsistent pg_dump options in the UI (CLUS-6624)

Maintenance Release: November 18th, 2025

  • Build:

    • clustercontrol-controller-2.3.4-17075
    • clustercontrol-mcc-2.3.4-654
    • clustercontrol2-2.3.4-2169

  • Controller:

    • Address an issue with the Galera donor node stuck in read-only mode after resync. An alarm will be triggered if the donor remains in read-only (CLUS-6566).
    • Address an issue with disabling read_only and super_read_only not taking effect with Percona XtraDB cluster (CLUS-6672).
    • Address an issue adding a replica when restoring a cloud backup using the default home directory instead of the configured backup mount point (CLUS-6586).
    • Address an UI issue with HAProxy advanced settings not correctly reflecting the correct RO/RW port used (CLUS-6670).

  • GUI:

    • Address an issue to remove ‘beta’ label from QM agents (CLUS-6682).
    • Address tooltips improvements for partial backup with pg_dump (CLUS-6630).

Maintenance Release: October 30th, 2025

  • Build:

    • clustercontrol-controller-2.3.4-16739

  • Controller:

    • Addressed hanged jobs occurring in the 2.3.4 release (CLUS-6606).
    • Added several handy scripts for improved usability.
    • Improved trace logging and applied various clang fixes (CLUS-6492).
    • Enhanced detection of cmon.service in controller_pool.sh (CLUS-6492).
    • Improved cluster monitoring traces in cmon.log (CLUS-6492).
    • Added thread info logging after instance creation (CLUS-6492).
    • Improved formatting and added traces for s9s_func_thread failures in cmon (CLUS-6492).
    • Removed controllers with duplicate hostname and port on start (CLUS-6492).
    • Forced cmonhostmanager cache update (reload from DB) when Galera has no servers (CLUS-6492).
    • Prevented m_galeraServers from being cleared when cache is empty (Fix for CLUS-6492).
    • Avoided updating cmonhostmanager to prevent deadlocks (CLUS-6492, CLUS-6606).
    • Fixed various clang-tidy issues (CLUS-6612, CLUS-6581).
    • PostgreSQL:
      • Improved logs during backup restore operations (CLUS-6612).
      • During pg_restore, directories are now processed correctly (CLUS-6581).
      • Prevented re-application of schema filters during backup restore (CLUS-6581).
      • PgDump: Use pg_dumpall restore when the DB name is * (CLUS-6612).
      • PgBackRest: Simplified filtering for verifyBackup mode (CLUS-6612).
      • PgBackRest: Filter backup candidates by the requested target host port (CLUS-6612).
      • Restore host now inherits port from the backup record (CLUS-6612).
      • PostgreSQL 18: Fixed upgrade process by adding --no-data-checksums when needed (CLUS-6522).

Maintenance Release: October 24th, 2025

  • Build:

    • clustercontrol-controller-2.3.4-16604

  • Controller:

    • Address an issue to update an expired repository GPG key for MySQL (CLUS-6604)

Maintenance Release: October 23rd, 2025

  • Build:

    • clustercontrol-controller-2.3.4-16591
    • clustercontrol-mcc-2.3.4-626
    • clustercontrol2-2.3.4-2151

  • Controller:

    • Address an issue with supported versions for TimescaleDB. Only available with PostgreSQL 16 (not 17 or 18 at the moment) (CLUS-6589)
    • Address an issue with node removal with Redis / Valkey cluster (CLUS-6464)
    • Address an issue when upgrading CMON 2.3.1 to 2.3.4 during DB migration (CLUS-6559)
    • Address an issue with pg_BackRest when the env MODULSHOME is used (CLUS-6562)
    • Address an issue with partial backup restore with pg_dump when using sudo user (CLUS-6538)
    • Address an improvement to support LDAP_OPT_REFERRALS with LDAP (CLUS-6355)
    • Address an issue with running CMON jobs being purged (CLUS-6520)
    • Address an improvement to support Redis 8 (CLUS-6496)
    • Address an improvement to support PostgreSQL 18 (CLUS-6497)
    • Limitations (will be addressed with a future patch):
      • Major upgrade from 17 to 18 is not currently supported
      • Logical Replication Deployments / Imports is not currently supported

  • GUI:

    • Enable TimescaleDB as an PostgreSQL action/job has been deprecated - not supported. Please deploy a TimescaleDB cluster or manually enable the extension if required (CLUS-6598)
    • Address an issue with the tooltip for pg_dumpall (CLUS-6582)
    • Address an issue with Query Monitor / Agents with Group Replication (CLUS-6525)
    • Address an issue with missing ‘bootstrap’ option when restoring MySQL backups (CLUS-6513)
    • Address an issue with PostgreSQL backup schedules where ‘verify backup’ option was not persisted/correctly set (CLUS-6535)
    • Address an issue with missing partial backup options when selecting ‘mariadb-dump’ as backup method (CLUS-6485)
    • Address an issue with a broken cluster list filter dialog (CLUS-6500)
    • Address minor cosmetic with the CMON pool wizard and UI (CLUS-6435, CLUS-6507, CLUS-6577)

Initial Release: October 7th, 2025

  • Build:
    • clustercontrol2-2.3.4-2138
    • clustercontrol-mcc-2.3.4-604
    • clustercontrol-controller-2.3.4-16304
    • clustercontrol-kuber-proxy-0.1.0-589
    • clustercontrol-proxy-2.3.4-129
    • clustercontrol-cloud-2.3.4-446
    • clustercontrol-clud-2.3.4-446
    • clustercontrol-notifications-2.3.4-388
    • clustercontrol-ssh-2.3.4-229
    • s9s-tools 1.9.2025100911

We’re pleased to announce the release of ClusterControl v2.3.4 which adds support for:

  • MySQL Group Replication for native multi-primary HA
  • Percona Server Pro 8.4
  • Backup and restore improvements specifically with external and partial backups for MySQL/PostgreSQL
  • Preview of a scalable CMON Controller pool

Features

MySQL Group Replication

Whether you’re building a high-availability service or planning for zero-downtime operations, Group Replication offers a robust MySQL-native solution that minimizes operational overhead while maximizing resilience:

  • Automatic failover with built-in consensus-based membership and leader election.
  • Synchronous replication using Paxos-like protocols to ensure strong consistency across nodes.
  • Multi-primary support for concurrent writes on multiple nodes with conflict detection.
  • Seamless scaling of reads and writes without manual topology management.
  • Self-healing clusters, with automatic member rejoin and state reconciliation.
  • Load balancing, with HAProxy and Keepalived for high availability routing.

Backup and Restore

  • Restore external backup
    • Various improvements / fixes for MySQL Replication & PostgreSQL
  • Partial Backup & Restore
    • Supported now with Percona Xtrabackup, MariaDB mariadb-backup and pg_dump

Scaling and High Availability - ClusterControl CMON Controller Pool

This update brings a preview of CMON controller pooling, dynamic cluster assignment, and Kubernetes-native configuration storage for K8s and non-K8s environments.

  • Controller Pooling: You can now run multiple collocated controllers (on the same or different hosts) with isolated configs and coordinated cluster ownership.
  • Cluster Auto-Assignment: Clusters are dynamically and transactionally assigned across active controllers. The system ensures no cluster is ever left orphaned. Ensures zero downtime and failover with minimal manual intervention
  • Graceful Scaling: When scaling your controller pool up/down, CC automatically redistributes clusters while preserving service continuity.

Miscellaneous

  • Support for Percona Server Pro 8.4
Copyright © 2014-2025 Severalnines AB. All rights reserved.

Severalnines and ClusterControl are registered trademarks in the US, UK, and EU. The 3rd-party trademarks on this site are property of their respective owners and are used for referential purposes only.

Made with Material for MkDocs