Skip to content

MariaDB/Galera

This page is the starting point for MariaDB/Galera, the mariadb role: the containers mariadb and mariadb_clustercheck, on the control nodes of the example inventory, selected by the tag mariadb. The lists below say where the study covers how to operate it, its impact, its dependencies and the evidence, what to check around a run and what stays open; the service card follows them.

Where to look

Operate

Impact and risk

Dependencies

Evidence

Before and after a run

  • Pre-execution checklist, step 3 (batch size): ANSIBLE_SERIAL covers reconfigure/upgrade and -e kolla_serial=N covers deploy; mariadb and rabbitmq ignore it.
  • Pre-execution checklist, step 5 (health gates): before the run, check Galera: wsrep_cluster_size, wsrep_local_state_comment=Synced. Repeat after each batch.
  • Pre-execution checklist, step 8 (rollback material): a MariaDB backup (kolla-ansible mariadb_backup).
  • Tags §3.1, row mariadb: safe on its own "only on a healthy, fully synced cluster"; prerequisite: every node Synced.
  • R9, precautions for a MariaDB config change or image change: check wsrep_cluster_size and Synced on all nodes before; take a backup (mariadb_backup); watch between waves.
  • R18, precautions for mariadb_recovery: only when the cluster is fully down; identify the node with the highest seqno manually first.
  • R22, inventory reordering (haproxy, mariadb, glance, nova-conductor, keystone groups): MariaDB restart waves and active backend move. Precautions: treat inventory order as configuration; never reorder casually.
  • fact sheet B §1.8, item 4 (fail-fast guards): reconfigure aborts if the cluster is fully stopped (multi-node) or any alive node is not Synced. kolla-ansible stop stops every node at once, so the next deploy/reconfigure fails until mariadb_recovery is run.

Known limits and open points

  • Known bugs and limitations, row mariadb: the restart waves have no any_errors_fatal; recovery stops healthy nodes; clustercheck started even when disabled. Status: CODE: a node that fails its WSREP wait marks only its own host failed, and the next wave still restarts its nodes (R9); whether the failed node is then still joining, and how large the quorum and availability risk is, NOT AVAILABLE.
  • Q6: "MariaDB rolling restart under load with a failure in wave 1: do later waves proceed?" Status: control flow answered from the source CODE: a node that fails its WSREP wait marks only its own host failed, and the next wave still restarts its nodes. NOT AVAILABLE Whether the failed node is then still joining, whether that can leave a single Synced node acting as donor, and so how large the quorum and availability risk is.
  • What is not known, row "MariaDB": the sources do not establish whether node 0's restart wave runs first; after a failed WSREP wait (the next wave still restarts its nodes CODE, R9), whether the failed node is then still joining, whether that can leave a single Synced node acting as donor, and how large the quorum and availability risk is; nor whether the DB containers come back as a non-primary cluster after a full stop, and whether a manual start left with --wsrep-new-cluster bootstraps a new cluster.
  • What is not known: the sources do not establish dumb-init signal forwarding for mysqld_safe under the MariaDB image's own entrypoint, whether xtrace writes one-shot secrets (the MariaDB root password among them) to the container's stderr, or the cost of the recursive chown on large trees (the MariaDB datadir among them).
  • fact sheet B §1.8, "Operationally notable behaviour, risks and oddities", eleven items: reconfigure with a galera.cnf change; clustercheck restart is not rolling; WSREP-sync wait is only about 60 s; fail-fast guards; recovery takes the whole cluster down; recovery bug; recovery assumptions; upgrade; backup; credentials on the command line; scale-out restarts the whole cluster.
  • Kolla images §6, items 10 and 16: recursive chown on data volumes at every start, the MariaDB datadir among them CODE; MariaDB has no automatic Galera recovery.
  • R20: a Glance reconfigure or deploy toggles the global MariaDB log_bin_trust_function_creators 1→0 around db sync.

Service card

The rows of the card are explained in How to read a service card.

Item Value
Play site.yml:471
Tags mariadb
Play host patterns mariadb
Node types in play (example inventory) control
Node types running its containers control
Containers mariadb, mariadb_clustercheck
deploy/reconfigure flow (tasks/deploy.yml) config → check-containers → bootstrap → flush handlers → register → check
Task files backup.yml, bootstrap.yml, bootstrap_cluster.yml, check-containers.yml, check.yml, config.yml, deploy-containers.yml, deploy.yml, loadbalancer.yml, lookup_cluster.yml, main.yml, precheck.yml, pull.yml, reconfigure.yml, recover_cluster.yml, register.yml, restart_services.yml, stop.yml, upgrade.yml, wait_for_loadbalancer.yml
Templates backup.my.cnf.j2, galera.cnf.j2, mariabackup.json.j2, mariadb-clustercheck.json.j2, mariadb.json.j2, wsrep-notify.sh.j2, xinetd-clustercheck.j2
Handlers (definition order) Starting first MariaDB container; Wait for first MariaDB service port liveness; Wait for first MariaDB service to sync WSREP; Creating haproxy mysql user; Restart MariaDB on existing cluster members; Start MariaDB on new nodes; Ensure MariaDB is running normally on bootstrap host; Restart mariadb-clustercheck container
Key variables enable_mariadb, mariadb_*, database_address, enable_external_mariadb_load_balancer
Role dependencies common (meta)