Using COC to Batch Execute Commands for Multiple ECSs

Scenarios

As Huawei Cloud ECSs have grown exponentially during the advanced stages of digital transformation, they play a critical role in supporting stable business systems. However, traditional O&M models now struggle to scale alongside massive resource pools that often contain hundreds or thousands of ECSs.

Before COC's batch execution was introduced, O&M teams typically relied on RDP or SSH to log in to ECSs individually. This approach works for limited resources, but it creates significant bottlenecks during high-frequency scenarios, such as batch version upgrades, large-scale configuration alignment, urgent security baseline hardening, and health check script delivery. These manual operations slow down application responses and make it difficult to maintain consistency as the number of ECSs increases.

In large-scale scenarios, traditional O&M methods are caught in a conflict between efficiency and risks.

Efficiency conflict: Individually vs. Batch
A direct contradiction exists between the linear time required for manual operations and the non-linear growth of business demands. For example, updating 10 servers may take only half an hour, but updating 100 servers may take several hours or even longer. This not only consumes a large number of human resources, but also causes a long maintenance window, reducing service availability.
Risk conflict: Human error vs. Consistency
When manually writing scripts or repeatedly entering commands, O&M personnel are prone to typos or parameter misconfigurations (such as incorrect paths or version numbers) due to fatigue. More critically, when different O&M personnel perform tasks at different times without a unified standard, the cluster suffers from "configuration drift." This means that although the hardware remains identical, the software environments diverge. This is a common root cause of production outages.

Solutions

Huawei Cloud COC's batch command execution builds a standardized, automated, and auditable O&M closed-loop system. It evolves traditional manual serial O&M into automated parallel O&M.

Replacing serial operations with batch execution: reducing costs and increasing efficiency

This feature allows you to easily filter ECSs by tag, enterprise project, or resource pool, reducing hours of manual labor into minutes. O&M personnel only need to code a standardized script once (using Shell, Python, or Bat). The system then uses UniAgent to dispatch tasks to all target ECSs simultaneously. This enables parallel processing for large-scale operations and drastically accelerates response times.
Mitigating risk through standardization: ensuring consistency

This feature enforces pre-execution script validation and OS isolation (ensuring separate execution for Linux and Windows). By using a unified script repository and standardized operational parameters, the system eliminates human error and inconsistency. This ensures that all instances within a cluster maintain a synchronized configuration at all times, fundamentally reducing configuration drift and lowering the rate of operation failures.

This section describes how to batch execute commands to query the disk usage of multiple ECSs.

Precautions

The target ECSs must be running.
The UniAgent of the ECSs is running. For details, see Installing UniAgent on a Host.
If the selected ECSs run on different OSs (Linux and Windows), execute commands for the ECSs separately.

Prerequisites

You have registered a Huawei ID and enabled Huawei Cloud services and logged in to the system.

Batch Executing Commands

Log in to COC.
In the navigation pane, choose Resource O&M > Batch Resource Operations.
On the Elastic Cloud Server tab page, click Execute Commands.

Set the parameters for batch command execution by referring to Table 1.

**Table 1** Parameters for batch command execution
Parameter	Example Value	Description
Target Instance	Target ECSs	Click Add and select the target ECSs.
Operating System	Linux	OS type of the target instances. Linux and Windows are supported. Select an OS type the same as that of the target instances.
Execution User	root	For ECSs running on Linux, the preset value is root. You can change it as needed.
Execution Plan	Execute Now	Execution plan of the command. The default value is Execute Now and cannot be changed.
Timeout Interval	60	A timeout interval for executing a command. If the time limit is reached, the system forcibly terminates the sending process. The unit is seconds.
Command Type	Shell	Type of the script command to be sent.
Command Input	df -h	The command must be able to return results after a single execution. Interactive commands requiring secondary input are not supported.