Calling RunCommandRequest every 60 seconds got 104 error (Connection aborted)

[tthnu]

Python version: 3.10.12
SDK-Version:2.16.0

I need to monitor the task process which run on ECS. So I called RunCommandRequest every 60 seconds to get the process info. But got 104 error:
HttpError occurred. Host:ecs-cn-hangzhou.aliyuncs.com SDK-Version:2.16.0 ClientException:SDK.HttpError ('Connection aborted.', ConnectionResetError(104, 'Connection reset by peer'))

Is there any limit to the frequency of RunCommandRequest calling?

[yndu13]

Hi,

Could you please let us know how many requests per second you are making to the API? Additionally, could you provide us with a usage example? This information will help us better understand the issue you are experiencing.

In the meantime, we recommend using a single global instance of AcsClient. When initializing AcsClient, you can configure the connection pool size to better suit your needs. For example:

acs_client = AcsClient(pool_size=128)
The pool_size value should be adjusted based on your specific use case and business requirements.

Please let us know if you need further assistance.

[tthnu]

Hi,

Could you please let us know how many requests per second you are making to the API? Additionally, could you provide us with a usage example? This information will help us better understand the issue you are experiencing.

In the meantime, we recommend using a single global instance of AcsClient. When initializing AcsClient, you can configure the connection pool size to better suit your needs. For example:

acs_client = AcsClient(pool_size=128) The pool_size value should be adjusted based on your specific use case and business requirements.

Please let us know if you need further assistance.

Thanks for your reply!

I call RunCommandRequest for ONE time every 60 seconds as I say in title. Below is my example code. It's almost a total copy of the code in interface user guide.

#### get client for one time ONLY at the begin of program 
client = AcsClient(args.ali_key, args.ali_secret, args.ali_ecs_region)


def run_command_async(client, cmdcontent, instance_id, timeout=60):
    try:
        request = RunCommandRequest()
        request.set_accept_format('json')

        request.set_Type("RunShellScript")
        request.set_CommandContent(cmdcontent)
        request.set_InstanceIds([instance_id])
        # 执行命令的超时时间，单位s,默认是60s,请根据执行的实际命令来设置
        request.set_Timeout(timeout)
        response = client.do_action_with_exception(request)
        # invoke_id = json.loads(response).get("InvokeId")
        # return invoke_id
        return json.loads(response)
    except Exception as e:
        return {
            'error': f'Running command error occurred: {str(e)}',
            'http_code': -1,
        }

def get_command_output(client, command_id, instance_id):
    """
    获取阿里云 ECS 命令执行的输出结果
    
    Args:
        command_id (str): 命令 ID
        instance_id (str): 实例 ID
    
    Returns:
        dict: 包含命令执行输出的字典
    """
    try:
        # 创建请求对象
        request = DescribeInvocationResultsRequest()
        request.set_accept_format('json')
        
        # 设置请求参数
        request.set_CommandId(command_id)
        request.set_InstanceId(instance_id)
        
        # 发送请求
        response = client.do_action_with_exception(request)
        result = json.loads(response)
        
        # 解析结果
        if result['Invocation']['InvocationResults']['InvocationResult']:
            invocation_result = result['Invocation']['InvocationResults']['InvocationResult'][0]
            
            # 输出是 Base64 编码的，需要解码
            if invocation_result['Output']:
                output = base64.b64decode(invocation_result['Output']).decode('utf-8')
            else:
                output = "No output available"
                
            return {
                'CommandId': command_id,
                'InstanceId': instance_id,
                'Status': invocation_result['InvocationStatus'],
                'Output': output,
                'ExitCode': invocation_result.get('ExitCode'),
                'ErrorInfo': invocation_result.get('ErrorInfo', '')
            }
        else:
            return {
                'error': 'No results found for the specified command and instance'
            }
            
    except Exception as e:
        return {
            'error': f'Running command error occurred: {str(e)}',
            'http_code': -1,
        }

### call RunCommandRequest and DescribeInvocationResultsRequest
res = run_command_async(client, main_cmd, instance_id)
output = get_command_output(client, command_id, instance_id)

time.sleep(60)

I think this frequency will not encounter any TPS limit.

BTW It's OK after using V2 SDK instead of V1 . I can do more test If you need more information to work it out.

from alibabacloud_ecs20140526.client import Client as EcsClient
from alibabacloud_ecs20140526 import models as ecs_models


    def run_command(
        client: EcsClient,
        content: str,
        instance_id: str,
        region_id: str,
    ) -> str:
        """
        RunCommand  执行脚本
        """
        req = ecs_models.RunCommandRequest()
        # ECS实例ID列表
        req.instance_id = [
            instance_id
        ]
        # 可用区ID
        req.region_id = region_id
        # 脚本的明文内容或者Base64编码后的内容
        req.command_content = content
        # 运维脚本的语言类型
        req.type = 'RunShellScript'
        resp = client.run_command(req)
        # ConsoleClient.log('--------------------执行脚本--------------------')
        # ConsoleClient.log(f'脚本ID {resp.body.command_id} 请求ID {resp.body.request_id}')
        return resp.body.command_id