3
votes

I'm trying to write a script using boto3 to start an instance and wait until it is started. As per the documentation of wait_until_running, it should wait until the instance is fully started (I"m assuming checks should be OK) but unfortunately it only works for wait_until_stopped and incase of wait_until_running it just starts the instance and doesn't wait until it is completely started. Not sure if I'm doing something wrong here or this is a bug of boto3.

Here is the code:

import boto3


ec2 = boto3.resource('ec2',region_name="ap-southeast-2")
ec2_id = 'i-xxxxxxxx'
instance = ec2.Instance(id=ec2_id)
print("starting instance " + ec2_id)
instance.start()
instance.wait_until_running()
print("instance started")
3
The waiter only waits until state is changed to running of the following: pending, running, shutting-down, terminated, stopping, stopped. - hjpotter92
wait_until_running() should wait until the instance's state has changed to running. It will not wait until the operating system has fully booted, and all services have started, etc. which is what I assume you mean when you say "completely started". - Mark B
Thaks @MarkB, I'm just thinking to put a sleep interval for a few seconds and check until port 22 is accessible (which makes sure that instance is finally up). Is this a good approach or is there any other better AWS way to get this done. - kolaman

3 Answers

7
votes

Thanks to @Mark B @Madhurya Gandi here is the solution that worked in my case:

import boto3,socket
retries = 10
retry_delay=10
retry_count = 0
ec2 = boto3.resource('ec2',region_name="ap-southeast-2")
ec2_id = 'i-xxxxxxxx'
instance = ec2.Instance(id=ec2_id)
print("starting instance " + ec2_id)
instance.start()
instance.wait_until_running()
while retry_count <= retries:
    sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    result = sock.connect_ex((instance.public_ip_address,22))
    if result == 0:
        Print "Instance is UP & accessible on port 22, the IP address is:  ",instance.public_ip_address
        break
    else:
        print "instance is still down retrying . . . "
        time.sleep(retry_delay)
   except ClientError as e:
       print('Error', e)
0
votes

I believe the right way to do it is as follows:

   instance.wait_until_running(
            Filters=[
                {
                    'Name': 'instance-state-name',
                    'Values': [
                        'running',
                    ]
                },
            ]
   )

tested and worked for me

-1
votes

I've tried instance.wait_until_running(). It took time to update the instance to running state. As per amazon docs link, it says that the instances take a minimum of 60seconds to spin up. Here's a sample code that worked for me. Hope it helps!

;to create 5 instances

ec2.create_instances(ImageId='<ami-image-id>', MinCount=1, MaxCount=5)
time.sleep(60)

;print your instances