Skip to main content

Troubleshooting — MariaDB

The cluster stays in "Creating" status​

Cause: the nodes and their volumes are being provisioned. This can take several minutes, longer with 3 or 5 replicas.

Solution:

  1. Wait a few minutes and refresh the cluster page.
  2. If the status does not change after about fifteen minutes, contact support, providing the project and the cluster name.

The cluster changes to "Error" status​

Cause: the platform could not apply the cluster's creation or its last change. After a change, the cluster may keep running with its previous configuration.

Solution: contact support, providing the project, the cluster name and the change requested.

Connection refused or timed out​

Cause: external access is disabled, the IP address has not been assigned yet, or the client is using the wrong address or port.

Solution:

  1. In the Connection and network card, check that the External zone contains an address, and not Disabled or "Waiting for allocation…". From a VM of the project, use the host in the Internal zone instead.
  2. Use port 3306 and test connectivity:
    mysqladmin -h <host> -P 3306 -u <user> -p ping
  3. Check that no outbound firewall on your network blocks port 3306.

Access denied for user​

Cause: wrong password or password revoked by a rotation, or user without rights on the specified database.

Solution:

  1. In the user list, check the Databases column: the user must have access to the database in use.
  2. Add the access via Actions → Manage Access if necessary.
  3. If in doubt about the password, generate a new one via Actions → Change Password and update your applications.

Error when adding an access or a user​

Cause: the database name or username does not follow the naming rules, or the access targets a system database (mysql, information_schema, performance_schema, sys).

Solution: use only lowercase letters, digits and hyphens, starting with a letter and ending with a letter or a digit. Underscores (_) and uppercase letters are not accepted. Grant access to an application database, never to a system database. See MariaDB concepts.

A mysql line is shown in error in "Manage Access"​

Cause: the user was created with the cluster creation wizard before console v1.3.0, which gave it a role on the mysql system database. A system database can no longer receive access, and updating the user is refused as long as this line is present.

Solution: remove the mysql line, add the user's application database if needed, then click Save. The "Access updated" message confirms the change.

Disk space full​

Cause: the data volume (including binary logs) has reached the Allocated Size.

Solution:

  1. Measure the space used per database:
    SELECT table_schema, ROUND(SUM(data_length + index_length) / 1024 / 1024, 1) AS size_mb
    FROM information_schema.tables
    GROUP BY table_schema;
  2. Increase the Disk size (GB) via Edit, within the project's storage quota. The size cannot be reduced afterwards. See Scale resources.
  3. Delete obsolete data, then optimize the tables concerned (OPTIMIZE TABLE).

Replication out of sync​

Cause: a replica can no longer keep up with the primary (high write load, insufficient resources, infrastructure incident).

Solution: resynchronizing a replica is not offered in the console. Contact support, providing the project and the cluster name.