AWS Lambda Python Boto3启动EC2实例遇UnboundLocalError问题咨询

问题原因 #

你的代码在启动EC2实例时触发UnboundLocalError，根源有两点：

1. 变量未定义场景：如果AWS账号下没有任何EC2实例，for each_scheduled_instance in ec2_console_resource.instances.all()循环根本不会执行，each_scheduled_instance变量从未被赋值，后续调用each_scheduled_instance.wait_until_running()自然会报错。
2. 逻辑缺陷：就算存在EC2实例，当前写法也只会等待最后一个启动的实例进入运行状态，其他实例的启动状态没有被监控。

报错信息 #

{
    "errorMessage": "local variable 'each_scheduled_instance' referenced before assignment",
    "errorType": "UnboundLocalError",
    "requestId": "3054209f-5cf7-429e-bcc7-7109a3b28a29",
    "stackTrace": [
        "  File \"/var/task/lambda_function.py\", line 19, in lambda_handler\n    return dealwithec2instances(event)\n",
        "  File \"/var/task/lambda_function.py\", line 25, in dealwithec2instances\n    return ec2_instances_start(event)\n",
        "  File \"/var/task/lambda_function.py\", line 40, in ec2_instances_start\n    each_scheduled_instance.wait_until_running()\n"
    ]
}

原代码 #

import boto3
import logging
import os
from pprint import pprint

logger = logging.getLogger()
logger.setLevel(logging.INFO)

def lambda_handler(event, context):
    varbotname = event['bot']['name']
    logger.info('bot name is ' +str(varbotname))
    return dealwithec2instances(event)

def dealwithec2instances(event):
    print('in dealwithec2instances')
    intentname = event['currentIntent']['name']
    if intentname == 'StartInstances':
        return ec2_instances_start(event)

def ec2_instances_start(event):
    print('in ec2 instances method')
    aws_mgmt_console = boto3.session.Session()
    ec2_console_resource = aws_mgmt_console.resource('ec2')
    ec2_console_client = aws_mgmt_console.client('ec2')

    waiter=ec2_console_client.get_waiter('instance_running')

    for each_scheduled_instance in  ec2_console_resource.instances.all():
        each_scheduled_instance.start()
    
    each_scheduled_instance.wait_until_running()    

    return 'success'

修复方案 #

方案一：逐个实例等待启动（资源API方式）

先检查是否存在实例，再逐个启动并等待：

import boto3
import logging
import os
from pprint import pprint

logger = logging.getLogger()
logger.setLevel(logging.INFO)

def lambda_handler(event, context):
    varbotname = event['bot']['name']
    logger.info('bot name is ' +str(varbotname))
    return dealwithec2instances(event)

def dealwithec2instances(event):
    print('in dealwithec2instances')
    intentname = event['currentIntent']['name']
    if intentname == 'StartInstances':
        return ec2_instances_start(event)

def ec2_instances_start(event):
    print('in ec2 instances method')
    aws_mgmt_console = boto3.session.Session()
    ec2_console_resource = aws_mgmt_console.resource('ec2')

    # 获取所有实例并转换为列表，方便检查是否为空
    instances = list(ec2_console_resource.instances.all())
    
    if not instances:
        logger.info("No EC2 instances found to start")
        return 'no_instances_found'

    # 逐个启动并等待实例进入运行状态
    for each_instance in instances:
        each_instance.start()
        each_instance.wait_until_running()    

    return 'success'

方案二：批量等待实例启动（客户端API方式，效率更高）

使用EC2客户端的waiter批量监控所有实例状态，适合大量实例场景：

import boto3
import logging
import os
from pprint import pprint

logger = logging.getLogger()
logger.setLevel(logging.INFO)

def lambda_handler(event, context):
    varbotname = event['bot']['name']
    logger.info('bot name is ' +str(varbotname))
    return dealwithec2instances(event)

def dealwithec2instances(event):
    print('in dealwithec2instances')
    intentname = event['currentIntent']['name']
    if intentname == 'StartInstances':
        return ec2_instances_start(event)

def ec2_instances_start(event):
    print('in ec2 instances method')
    aws_mgmt_console = boto3.session.Session()
    ec2_console_client = aws_mgmt_console.client('ec2')

    # 收集所有实例ID
    instance_ids = []
    response = ec2_console_client.describe_instances()
    for reservation in response['Reservations']:
        for instance in reservation['Instances']:
            instance_ids.append(instance['InstanceId'])
    
    if not instance_ids:
        logger.info("No EC2 instances found to start")
        return 'no_instances_found'

    # 批量启动实例
    ec2_console_client.start_instances(InstanceIds=instance_ids)
    
    # 使用waiter等待所有实例进入运行状态
    waiter = ec2_console_client.get_waiter('instance_running')
    waiter.wait(InstanceIds=instance_ids)

    return 'success'

修改说明 #

• 新增实例存在性检查，避免无实例时变量未定义的错误
• 方案一保持资源API的简洁性，适合少量实例；方案二用客户端API批量操作，效率更高
• 确保所有启动的实例都被监控到运行状态，修复原逻辑只等待最后一个实例的缺陷

内容的提问来源于stack exchange，提问作者Jason