apache/dolphinscheduler · error · ServiceException

120034

120034

Error message

this cluster has been used in namespace,so you can't delete it.

What it means

ClusterServiceImpl.deleteClusterByCode refuses to delete a cluster that is still referenced by one or more K8s namespaces. Before deleting, it counts namespaces linked to the cluster code via k8sNamespaceDao.countByClusterCode(code) and throws ServiceException(Status.DELETE_CLUSTER_RELATED_NAMESPACE_EXISTS, code 120034) when the count is > 0. This is a referential-integrity guard: deleting the cluster would leave orphaned namespace relations.

Source

Thrown at dolphinscheduler-api/src/main/java/org/apache/dolphinscheduler/api/service/impl/ClusterServiceImpl.java:199

        return dto;
    }

    /**
     * delete cluster
     *
     * @param loginUser login user
     * @param code      cluster code
     */
    @Override
    public void deleteClusterByCode(User loginUser, Long code) {
        if (isNotAdmin(loginUser)) {
            throw new ServiceException(Status.USER_NO_OPERATION_PERM);
        }

        long relatedNamespaceNumber = k8sNamespaceDao.countByClusterCode(code);

        if (relatedNamespaceNumber > 0) {
            throw new ServiceException(Status.DELETE_CLUSTER_RELATED_NAMESPACE_EXISTS);
        }

        if (clusterDao.deleteByCode(code)) {
            return;
        }
        throw new ServiceException(Status.DELETE_CLUSTER_ERROR);
    }

    /**
     * update cluster
     *
     * @param loginUser login user
     * @param code      cluster code
     * @param name      cluster name
     * @param config    cluster config
     * @param desc      cluster desc
     */
    @Override

View on GitHub (pinned to 02eac45a1b)

Solutions

1. Delete all namespaces associated with the cluster first (K8s namespace management page or k8sNamespaceDao.deleteByClusterCode), then retry the cluster delete.
2. Verify which namespaces reference the cluster by querying the k8s namespace table/list API filtered by cluster code.
3. If the namespace rows are stale/orphaned (namespace no longer exists in K8s), remove them manually from the database (t_ds_k8s_namespace) and retry.
4. Only after the related namespace count reaches 0 will deleteClusterByCode succeed.

Example fix

// before: delete cluster directly
clusterService.deleteClusterByCode(loginUser, clusterCode);
// after: clean up related namespaces first
List<K8sNamespace> ns = k8sNamespaceDao.queryByClusterCode(clusterCode);
if (!ns.isEmpty()) {
    throw new IllegalStateException("Delete related namespaces first: " + ns.size());
}
clusterService.deleteClusterByCode(loginUser, clusterCode);

// Java
long nsCount = k8sNamespaceDao.countByClusterCode(clusterCode);
if (nsCount > 0) {
    throw new IllegalStateException("Cluster " + clusterCode + " still has " + nsCount + " related namespaces");
}

try {
    clusterService.deleteClusterByCode(loginUser, code);
} catch (ServiceException e) {
    if (e.getCode() == 120034) {
        // prompt user to delete related namespaces first
    } else {
        throw e;
    }
}

When it happens

Trigger: Calling DELETE on the cluster API endpoint (or ClusterService.deleteClusterByCode) for a cluster whose code appears in the t_ds_k8s_namespace table. Happens even when the namespaces themselves are stale or the cluster is no longer actually running in K8s.

Common situations: Admins tearing down an environment: they remove a K8s cluster registration in the UI while namespace records created for that cluster (e.g. during workflow execution with K8s task type) were never cleaned up. Also occurs after partial deletes or when re-registering a cluster under a new code but leaving old namespace rows.

Related errors

• 120025
• DELETE_ENVIRONMENT_RELATED_TASK_EXISTS
• 110012
• 1400004
• 120033

AI-assisted analysis of apache/dolphinscheduler@02eac45a1b (2026-09-06). Data as JSON: /api/errors/d1ef595def4d1ff1. Report an issue: GitHub.