Start, stop, or restart instances

Stay organized with collections Save and categorize content based on your preferences.

This page describes how to start, stop, or restart the primary instance and a read pool instance. You can also restart a secondary instance and a node in a read pool of an AlloyDB cluster.

The start, stop, or restart operations don't change the inbound public IP or the Private Services Connect URI of the instance.

Before you begin

• The Google Cloud project you are using must be enabled to access AlloyDB.

• You must have one of the following IAM roles in the Google Cloud project you are using:

  • (roles/alloydb.admin) (the AlloyDB Admin predefined IAM role)
  • (roles/owner) (the Owner basic IAM role)
  • (roles/editor) (the Editor basic IAM role)

  If you don't have any of these roles, contact your Organization Administrator to request access.

• The AlloyDB cluster is in READY state.

Activation policy

When you start or stop instance, the activation policy must be set to indicate the operation you are performing. The activation policy indicates if the instance is activated to accept connection requests.

Activation policy options are:

• ALWAYS - The instance is always up and running.
• NEVER- The instance is inactive and isn't accepting connections.

To start your instance, set your activation policy to ALWAYS. To stop your instance and avoid instance charges, set your activation policy to NEVER.

Stop an instance

You can manually stop a primary or read pool instance when needed. When you stop an instance, it remains stopped, and doesn't respond to application connections, until you start it again. You can't stop a secondary instance.

Stopping an instance temporarily halts automatic updates. The data stored in the cluster is unaffected and any backup configuration for the cluster remains active.

Stopped instances don't incur any compute charges associated with the instance. If public IP is enabled on the instance, then you incur idle static external IP charges.

To stop an instance:

gcloud alloydb instances update INSTANCE_ID \
 --region=REGION_ID \
 --cluster=CLUSTER_ID \
 --activation-policy=NEVER
 ```
Replace the following:

Start an instance

When you start the instance, AlloyDB applies the latest minor version for the database. As starting an instance isn't a maintenance activity, any maintenance deny periods configured for the instance are ignored.

To start a stopped instance:

gcloud alloydb instances update INSTANCE_ID \
 --region=REGION_ID \
 --cluster=CLUSTER_ID \
 --activation-policy=ALWAYS
 ```
Replace the following:

Restart an instance

When you initiate a restart of an instance, it disrupts all connections that are available on the instance. After the instance restarts, and it's ready to accept new connections. Restarting an instance doesn't change the private IP address of the instance.

Restart a primary,secondary, or read pool instance

gcloud alloydb instances restart INSTANCE_ID \
    --cluster=CLUSTER_ID \
    --region=REGION_ID \
    --project=PROJECT_ID \
    [--async]
    

Restart specific nodes of a read pool

You can use the gcloud CLI to restart one or more nodes of a read pool.

The following command shows how you can restart a node of a read pool:

gcloud alloydb instances restart INSTANCE_ID \
    --cluster=CLUSTER_ID \
    --region=REGION_ID \
    --node-ids=NODE_ID \
    

• INSTANCE_ID: the ID of the readpool instance.
• CLUSTER_ID: the ID of the cluster where the instance is available.
• NODE_ID: the ID of the node that needs a restart, for example-j71a-read-pool-054b8861-2ln8.
• PROJECT_ID: the ID of the project where the cluster is placed.
• --async: an optional flag to return immediately, without waiting for the operation in progress to complete.

The following example shows how you can restart multiple nodes:

gcloud alloydb instances restart INSTANCE_ID \
    --cluster=CLUSTER_ID \
    --region=REGION_ID \
    --node-ids=NODE_ID1,NODE_ID2 \
    

• INSTANCE_ID: the ID of the readpool instance.
• CLUSTER_ID: the ID of the cluster where the instance is available.
• NODE_ID1,NODE_ID2: the list of nodes that need a restart.
• PROJECT_ID: the ID of the project where the cluster is placed.
• --async: an optional flag to return immediately, without waiting for the operation in progress to complete.

What's next

• Learn about creating a primary instance
• Update Cluster and instance settings