Changing DCS Instance Specifications

Notes and Constraints

• It is recommended that you change instance specifications during off-peak hours. The specifications change may fail during peak hours, for example, when the memory usage or maximum CPU usage of an instance exceeds 90% or the write traffic surges. If the change fails, try again during off-peak hours.
• This function is unavailable on the console for legacy instances. In this case, submit a service ticket and contact the customer service to upgrade the instances to the latest version. After the upgrade, the instance specifications can be changed.
• Cluster DCS Redis 3.0 instances cannot be vertically scaled.
• If the reserved memory of a DCS Redis 3.0 or Memcached instance is insufficient, specifications scale-up may fail when the memory is used up. For details, see Reserved Memory.
• The number of replicas and the capacity must be changed separately.
• Only one replica can be deleted at a time.

Billing

Changing the type or specifications of a DCS instance will affect your billing. Before confirming the change, review the new price.

Video Tutorial

This video demonstrates how to change the specifications of a DCS instance. As product functions evolve, the GUI may vary. This tutorial is for reference only.

Instance Change Notice

For notice on changing the instance type, see Changing the Instance Type. For notice on changing the instance specifications, see Changing Instance Specifications.

Changing an Instance

1. Log in to the DCS console.
2. Click in the upper left corner of the console and select the region where your instance is located.
3. In the navigation pane, choose Cache Manager.
4. Choose More > Modify Specifications of the Operation column in the row containing the DCS instance.
5. On the Modify Specifications page, select the desired specification.
6. Set Apply Change to Now or During maintenance.

   Select During maintenance if the change will interrupt client connections.

   Table 4 Scenarios where specifications change will interrupt connections

   Specifications Change Task	When Connections Are Interrupted
   Scaling up a single-node or master/standby instance	Memory is increased from a size smaller than 8 GB to 8 GB or larger.
   Scaling down a Proxy Cluster or Redis Cluster instance	The number of shards is decreased.
   Changing the instance type	The instance type is changed between master/standby or read/write splitting and Proxy Cluster.
   Deleting replicas	Replicas are deleted from a master/standby, Redis Cluster, or read/write splitting instance.

   • If the specifications change does not interrupt client connections, the change will be performed immediately even if you select During maintenance.
   • After you submit a specifications change request, the request cannot be canceled. You can change the maintenance window to reschedule the change for a maximum of three times.
   • To cancel the change before the maintenance window elapses, submit a service ticket and contact the customer service to enable the change cancellation feature. After the feature is enabled, click Cancel next to the task on the Background Tasks page.
   • DCS Redis 3.0 and Memcached instances can only be changed immediately.
   • If you apply the change during maintenance, the change starts at any time within the maintenance window, rather than at the start time of the window.
   • If a large amount of data needs to be migrated when you scale down a Cluster instance, the operation may not be completed within the maintenance window.

7. Click Next. Confirm the change details and view the risk check results.
   • If any risk is found in the check, the instance may fail to be changed. In this case, rectify the fault by referring to Table 5.
   • If the check result is normal, the checked items have no change failure risks.
   • If the check fails, the possible causes are as follows:
     • Failed to connect to the master node of the instance. In this case, check the instance status.
     • A system exception occurred. In this case, click Check Again later.
   • Click Stop Check to stop the check. Click Check Again to restart the check.
   • If you want to proceed with the change after manually stopping the check or if the check result reports an exception, select I understand the risks.

     Table 5 Risk check items

     Check Item	Reason for Check	Solution
     Non-standard configuration check NOTE: Currently, non-standard configuration check is available only in certain regions, such as CN North-Beijing4, CN East-Shanghai1, and CN East-Shanghai2. The supported regions are provided on the console. The following non-standard check items are supported: Non-standard bandwidth of a single node Non-standard memory of a single node Non-standard number of replicas on a Redis Cluster instance Non-standard number of proxy nodes on a Proxy Cluster instance Non-standard maxclients (exceeding the maximum number of allowed connections) on a Proxy Cluster instance	If your instance has non-standard configurations, the console displays a message indicating that they will be converted to standard configurations during the change. Among these non-standard configurations, you can retain only the non-standard bandwidth or number of proxy nodes in the change.	If your instance does not have non-standard configurations, the check result is normal and no action is required. If the instance has non-standard configurations, determine whether to proceed with the change or whether to retain the non-standard bandwidth or number of proxy nodes.
     Node status	Abnormal node status will lead to instance change failures.	If the node status is abnormal, submit a service ticket and contact the customer service.
     Dataset memory distribution check (This check item applies only to Proxy Cluster and Redis Cluster instances.)	During the specifications change of a Redis Cluster instance, data is migrated between nodes. Redis has a limit on key migration. If the instance has any single key greater than 512 MB, specifications change will fail when big key migration between nodes times out. If the instance dataset memory is unevenly distributed among nodes and the difference is greater than 512 MB, the instance has a big key and the change may fail.	If there are big keys on your instance, Handle big keys before proceeding with the change.
     Memory usage check	If the memory usage of a node is greater than 90%, keys may be evicted or the change may fail.	If the memory usage is too high, optimize the memory by optimizing big keys, scanning for expired keys, or deleting some keys.
     Network input traffic check (This check item applies only to single-node, read/write splitting, and master/standby instances.)	The change may fail if the network input traffic is too heavy and the write buffer overflows.	Perform the change during off-peak hours.
     CPU usage check	If the node CPU usage within the last 5 minutes is greater than 90%, the change may fail.	Perform the change during off-peak hours. Rectify the fault by referring to Troubleshooting High CPU Usage of a DCS Redis Instance.
     Resource capacity (This check item applies only to the scale-up of the shard capacity on a Cluster instance.)	During the scale-up of the shard capacity, if the VM resources of the instance are insufficient, instance nodes will be migrated. During the migration, service connections will be intermittently interrupted and the nodes will become read-only.	If there are risks in the resource capacity check, ensure that your client application has a reconnection mechanism and can handle exceptions. Otherwise, you may need to restart the client application after the change.

8. After the risk check is complete, click Next. After the change request is submitted, you can go to the Background Tasks page to view the change status.

   Specifications change of a single-node, master/standby, or read/write splitting DCS instance takes about 5 to 30 minutes to complete, while that of a Cluster instance takes longer.

   Click the task name on the Background Tasks page to view task details. After the specifications of an instance is successfully changed, its status changes to Running.

   Figure 1 Viewing background task details
   
   • If the specifications change of a single-node instance fails, the instance is temporarily unavailable and its specifications remain unchanged. Some management operations (such as parameter configuration and specifications change) are temporarily unavailable. After the specifications change is complete in the background, the instance changes to the new specifications and becomes available again.
   • If the specifications change of a master/standby or Cluster instance fails, the instance is still available for use with its original specifications. Some management operations (such as parameter configuration, backup, restoration, and specifications change) are temporarily unavailable. Remember not to read or write more data than allowed by the original specifications. Otherwise, data loss may occur.

Helpful Links

• To change instance specifications by calling APIs, see Modifying Instance Specifications.
• Can I Expand a Single Shard of a Cluster Instance (Scale-Up)?
• How Do I Add Shards to a Cluster DCS Redis Instance Without Changing the Memory?
• How Do I Handle an Error When I Use Lettuce to Connect to a Redis Cluster Instance After Specification Modification?
• Are DCS Instances Stopped or Restarted During Specification Modification?