jjc1138 / ec2start Goto Github PK

• accept comma separated list of types in argument
• choose cheapest, with first specified breaking any ties

https://aws.amazon.com/blogs/aws/amazon-ec2-update-streamlined-access-to-spot-capacity-smooth-price-changes-instance-hibernation/

• instead of deleting all rules, only delete the rule which is for the SSH/RDP port from a single IP address
• add an option to skip messing with the security group at all (the group to use will still need to be specified for spot instances)
  • maybe that should be the default?

Copy the logic from JJC1138/route53dynamicdns for upserting or deleting records depending on whether the instance has an IPv6 address assigned.

1. would make it easier for users to find instance in their console etc.
2. would make it possible to reconnect to a running spot instance using the other mode of ec2start if, for example, your client IP has changed so that you need to be reauthorized to connect in the security group

Mainly just to add the help text

botocore.exceptions.ClientError: An error occurred (InvalidParameterValue) when calling the DescribeSecurityGroups operation: Invalid value 'Utility' for groupName. You may not reference Amazon VPC security groups by name. Please use the corresponding id for this operation.

Grab the console log and look for the section like:

<14>Sep 18 18:13:01 ec2: #############################################################
<14>Sep 18 18:13:01 ec2: -----BEGIN SSH HOST KEY FINGERPRINTS-----
<14>Sep 18 18:13:01 ec2: 1024 SHA256:cmrwEeQz9q5EK7k9xXaN3Rc6OLLf05iYbd0hgCfIAN4 root@ip-172-31-16-118 (DSA)
<14>Sep 18 18:13:01 ec2: 256 SHA256:0b/g3YU9XAjw6ynBxfZtm+1B4l+JReUjls3/b46BkV8 root@ip-172-31-16-118 (ECDSA)
<14>Sep 18 18:13:01 ec2: 256 SHA256:2UgB9LxtU/HelSjHt3nUzDZQ3tSDauQRNrP28IDKtKw root@ip-172-31-16-118 (ED25519)
<14>Sep 18 18:13:01 ec2: 2048 SHA256:pK0qTBTwpQCd0V/asdmkhZogI6TLgkubYKXo9LVovbU root@ip-172-31-16-118 (RSA)
<14>Sep 18 18:13:01 ec2: -----END SSH HOST KEY FINGERPRINTS-----
<14>Sep 18 18:13:01 ec2: #############################################################

Parse the lines and add them with the hostname to our known_hosts.
Handle the console log being absent, which may happen when it's not the instance's first boot and it may be blank for a while after first starting too.
Do this stuff only if we don't have an entry for the server in our known_hosts, and if we don't then keep trying to get the console log until it works.

It should create the new image, optionally delete the old one and the corresponding snapshot, and optionally terminate the old instance.