Changing vCPUs and Memory of a GeminiDB Redis Instance

Scenarios

You can change instance vCPUs and memory on the console to accommodate your workloads. If your instance is overloaded, you are advised to add compute nodes.

Usage Notes

Nodes will be restarted one by one. When a node is being restarted, workloads on it are taken over by other nodes in the cluster. To ensure a smooth upgrade, check the connections and CPU load on the remaining nodes in advance to avoid latency increase and service exceptions caused by too many connections or heavy load. Each time a node is restarted, workloads will be interrupted for 5s to 10s. Install patches during off-peak hours to prevent instance overload. Note: When instance restoration is complete, its workloads may still be unavailable. You are advised to configure automated reconnection on the client so it can quickly connect to the node after a restart.

Billing

**Table 1** Billing
Billing Mode	Operation	Impact on Price
Yearly/Monthly	Specification upgrade	The new specifications will take effect within the remaining usage period. As the new configuration is priced higher than the old one, you will be charged the price difference. Example (The specifications and prices below are for reference only. The actual prices are displayed on the console.) Suppose you purchased a one-month, 3-node GeminiDB Redis instance with 1 vCPU and 6 GB of memory on April 8, 2023 and changed its specifications to 2 vCPUs \| 12 GB on April 18, 2023. The price for the old specifications is $941.70 USD/month, and that for the new specifications is $1883.40 USD/month. The price difference is calculated as follows: Price difference = Price for the new instance specifications × Remaining period – Price for the original instance specifications × Remaining period The remaining period in the formula is calculated by dividing remaining days in each calendar month by the total days in that month. In this example, the remaining period is 0.6581 [12 (remaining days in April)/30 (total days in April) + 8 (remaining days in May)/31 (total days in May)]. The price for specification upgrade is therefore $619.73 USD (1883.4 × 0.6581 – 941.7 × 0.6581). For details, see Yearly/Monthly Billing.
Yearly/Monthly	Specification downgrade	The new specifications will take effect within the remaining usage period. As the new configuration costs less than the previous one, Huawei Cloud will refund you the price difference. Example (The specifications and prices below are for reference only. The actual prices are displayed on the console.) Suppose you purchased a one-month, 3-node GeminiDB Redis instance with 2 vCPUs and 12 GB of memory on April 8, 2023 and changed its specifications to 1 vCPU \| 6 GB on April 18, 2023. The price for the old specifications is $1883.40 USD/month, and that for the new specifications is $941.70 USD/month. The price difference is calculated as follows: Price difference = Price for the original instance specifications × Remaining period – Price for the new instance specifications × Remaining period The remaining period in the formula is calculated by dividing remaining days in each calendar month by the total days in that month. In this example, the remaining period is 0.6581 [12 (remaining days in April)/30 (total days in April) + 8 (remaining days in May)/31 (total days in May)]. The refund amount is therefore $619.73 USD (1883.4 × 0.6581 – 941.7 × 0.6581). For details, see Yearly/Monthly Billing.
Pay-per-use	Specification upgrade	After the specifications are changed, a new order is generated, and you will be billed based on the new price. The old order automatically becomes invalid. For details, see Pay-per-use Billing.
Pay-per-use	Specification downgrade

Video Tutorial

This video shows how to change instance vCPUs and memory.

Procedure

Log in to the GeminiDB console.
In the navigation pane, choose Instances. On the displayed page, locate the required instance. In the Operation column, click Change Specifications or choose More > Change Specifications.

Figure 1 Changing specifications

In the DB Information area on the Basic Information page, click Change under Node Specifications.
Figure 2 Changing specifications
On the displayed page, select a specification change mode and required specifications, and click Next.
- Online change: During the change, instance nodes are upgraded in rolling mode, which has the minimal impact on services. The change duration is positively related to the number of nodes. Each node takes about 5 to 10 minutes. If there are a large number of nodes, wait patiently.
- Offline change: During offline change, all nodes are changed concurrently, which interrupts services for 10 to 20 minutes. Exercise caution when performing this operation. For your online production services, you are advised to perform the change online.
Figure 3 Changing specifications
On the displayed page, confirm the specifications.
- If you need to modify your settings, click Previous.
- If you do not need to modify your settings, click Submit.
View the change results.

In the DB Information area on the Basic Information page, you can see the new specifications.

Follow-up Operations

Return to the instance list. In the navigation pane, choose Task Center and check the progress of the change task.

If you have selected Immediate for Scheduled Time:
On the Instant Tasks tab, search for the task and check the execution progress. Instant tasks cannot be canceled.
If you have selected Duration maintenance window for Scheduled Time:
On the Scheduled Tasks page, search for the task and check the execution progress.

If the task is in the To be executed state, you can click Cancel to cancel the task.