meltwater/terraform-aws-asg-dns-handler

IndexError: list index out of range

Chilinot opened this issue · 4 comments

Chilinot commented

The lambda function is unable to create new Route53 records for launching instances and kills them over and over. I had to remove the lifecycle hooks to get it to stop killing my servers.

This is the error in the log that is printed when it fails to create a Route53 record for a launching instance:

list index out of range: IndexError
Traceback (most recent call last):
  File "/var/task/autoscale.py", line 138, in lambda_handler
    process_record(record)
  File "/var/task/autoscale.py", line 130, in process_record
    process_message(json.loads(record['Sns']['Message']))
  File "/var/task/autoscale.py", line 120, in process_message
    private_ip = fetch_private_ip_from_ec2(instance_id)
  File "/var/task/autoscale.py", line 23, in fetch_private_ip_from_ec2
    ip_address = ec2_response['Reservations'][0]['Instances'][0]['NetworkInterfaces'][0]['PrivateIpAddress']
IndexError: list index out of range
jimsheldon commented

Hello @Chilinot, sorry for the delayed response on this.

We will see if we can reproduce this. If you can provide any more details about your configuration, that might be helpful. Thanks

jimsheldon commented

@Chilinot were you trying to use this module for an autoscaling group that brought up instances with public IPs? I think that would explain the error you saw. Public IP support was added in #26 which is available in version 2.1.0.

If this is true, please give it a try and report back.

Chilinot commented

Hi, unfortunately we are no longer using this module. I remember i solved the issue somehow, however, chose to go with a different solution in the end. I'm closing this ticket.

Chilinot commented

Remembering now, i believe i had done some testing before deployments, and there were already some residual Route53 records from my testing. This made things go out of phase and broke the lambda script. I believe removing all clashing records and then setting this up again from the beginning solve the issue.

Hopefully it might help someone else that stumbles upon the issue.