Checking the Cluster Status

On the Clusters page of the GaussDB(DWS) management console, you can view the general information about a cluster in the cluster list, such as the cluster status, task information, recent events, and node flavor.

Querying General Information of a Cluster

Log in to the GaussDB(DWS) management console. In the navigation tree on the left, click Clusters. The cluster list displays all clusters. If there are a large number of clusters, you can turn pages to view the clusters in any status.

Enter the cluster name in the search box, and click image1 to search for a cluster. Alternatively, in the All projects drop-down list above the cluster list, select the target project. Click image2 to refresh the cluster list and billing mode. You can also click Search by Tag to search for clusters based on the criteria. For details, see Searching for Clusters Based on Tags.

Clusters are listed in chronological order by default, with the most recent clusters displayed at the top. Table 1 describes the cluster list parameters.

Table 1 Cluster list parameters

Parameter

Description

Cluster Name

Cluster name specified when a cluster is created.

Note

After a cluster is created, its name cannot be changed.

Cluster Status

Cluster running status. For details, see Cluster Status.

Task Information

Cluster task status. For details, see Cluster Task Information.

Node Flavor

Node flavors of clusters.

Enterprise Project

Enterprise project to which a cluster belongs.

Operation

Cluster Status

Table 2 Cluster status description

Status

Description

Available

Indicates that the cluster runs properly.

Read-only

A cluster goes into this state when the disk usage of the cluster or a single node in the cluster is greater than 90%. The cluster can still work in this state but supports only query operations. Write operations are not supported. When the cluster status becomes read-only, contact technical support engineers.

After the read-only status is canceled for the cluster, you are advised to perform the following operations:

  • Use the SQL client tool to connect to the database as the administrator and run the following command to periodically clear and reclaim the storage space:

    VACUUM FULL;
    

    After you delete data stored in GaussDB(DWS) data warehouses, dirty data may be generated possibly because the disk space is not released. This results in disk space waste. It is recommended that the storage space be cleared periodically.

  • You are advised to check the disk capacity and analyze whether the existing cluster specifications meet service requirements. If not, expand the cluster capacity. For details, see Scaling Out a Cluster.

Unbalanced

If the role of a GTM or DN in the cluster is different from the initial role, the cluster is in the Unbalanced state. In the Unbalanced state, the number of primary instances on some nodes increases. As a result, the load pressure is high. In this case, the cluster is normal, but the overall performance is not as good as that in a balanced state. You are advised to switch a cluster to the Available state during off-peak hours.

Redistributing

A cluster goes into this state when it detects that the service data on the original nodes is significantly larger than that on the new node after a new node is added to the cluster. In this case, the system automatically redistributes data on all nodes. The cluster can still work in this state.

Redistribution failed

A cluster goes into this state when data redistribution fails, but no data loss occurs. The cluster can still work in this state. You are advised to contact technical support.

Degraded

A cluster goes into this state when some nodes in the cluster are faulty, but the whole cluster runs properly. You are advised to contact technical support.

Unavailable

A cluster goes into this state when it cannot provide database services. You are advised to contact technical support.

Creating

A cluster goes into this state when it is being created.

Creation failed

A cluster goes into this state when it fails to be created.

Creating, restoring

A cluster goes into this state when it is being restored from a snapshot.

Cluster Task Information

Table 3 Task information description

Status

Description

Creating snapshot

Indicates that a snapshot is being created in the cluster.

Snapshot creation failed

Indicates that a snapshot fails to be created.

Configuring

Indicates that the system is storing modifications of cluster parameters.

Restarting

Indicates that a cluster is being restarted.

Restart failed

Indicates that a cluster fails to be restarted.

Scaling out

Indicates that a cluster is being scaled out.

Scale-out failed

Indicates that a cluster fails to be scaled out.

Expanding disk capacity

Indicates that disk capacity is being expanded.

Disk expansion failed

Indicates that disk capacity fails to be expanded.

Associating ELB

Indicates that ELB is being associated.

Failed to associate ELB

Indicates that ELB fails to be associated.

Disassociating ELB

Indicates that ELB is being disassociated.

Failed to disassociate ELB

Indicates that ELB fails to be disassociated.

Checking scale-in

The service is checking whether a cluster can be scaled in.

Scale-in check failed

A cluster does not meet the scale-in requirements. For example:

  • The value of default_storage_nodegroup is not installation.

  • In the cluster database, data_redis is a reserved redistribution schema, but the schema contains user tables.

  • The cluster disk space does not meet the scale-in requirements. For details, see Scaling In a Cluster.

Scaling in

A cluster is being scaled in.

Scale-in failed

The cluster scale-in fails. You need to manually scale in the cluster again as soon as possible, or your services will be affected.

Switching back

The primary/standby relationship of a cluster is being restored.

Switchback failed

The primary/standby relationship of a cluster fails to be restored. Possible causes are as follows.

  • Redo operations are being performed on DNs. Wait until the operations are completed and try again.

  • Failed to query DN redo information. Check tenant logs to identify the failure cause.

  • Primary/standby catchup is in progress. Wait until it is completed and try again.

  • Failed to query primary/standby catchup information. Check tenant logs to identify the failure cause.

  • Primary/standby catchup failed. Contact technical support or try again later. Check tenant logs to identify the failure cause.

  • The cluster is abnormal.

Changing node flavor

The cluster is being scaled.

Node flavor change failed

All specifications change failed

Changing all specifications

All the specifications of the cluster being changed.

All specifications change failed

Specifications change failed because of insufficient quotas or permissions, or abnormal cluster status.