sshkey | Go package for loading OpenSSH keys

You can use Data Source called aws_security_group to get details of an existing SG:

Properly quoting something isn't always just a matter of putting a quote at the beginning, and another at the end; if the string contains characters that'll have special meaning within that type of quotes, you need to escape them or otherwise disable that special meaning.

In this case, it's the $ (in print $7) that's causing trouble. Inside double-quotes, $ initiates parameter, command, etc substitution, so the shell replaces $7 with the value of the seventh argument to the current shell, and there isn't one, so you get ... | awk 'END{print }' being passed to ssh as the command to send to the remote computer. Note that while the $ is also in single-quotes, those are inside the double-quotes and hence get ignored (until they get to the remote shell, that is).

But there's also some confusion about case 1, the one that works. Because the command string is not quoted, and contains |, the local shell parses it into two separate commands in a pipeline:

Seems like these 2 documentation links are a lie:

• https://docs.okd.io/4.9/installing/installing_sno/install-sno-preparing-to-install-sno.html
• https://docs.okd.io/4.9/installing/installing_sno/install-sno-installing-sno.html

According to these 2 issues:

• https://github.com/openshift/okd/discussions/1012
• https://github.com/openshift/openshift-docs/issues/39759

OKD does not support "installation with Assisted Installer" and these links are "installation with Assisted Installer". Nice waste of time.

The problem was the special environment of Hetzner. I had to log into the nodes as soon as they started and reconfigure the hostname. After the next reboot everything was fine.

This is one of those things that in my opinion should not be ported directly to Ruby. While the OpenSSH binaries and source are routinely audited, lightly- or rarely-used Ruby gems or FFI wrappers wouldn't get the same level of scrutiny. Instead, you should use the Kernel#system or Kernel#` calls or the %x() subshell literal, depending on your use case.

For example:

I don't think you can have the iso variable set while using the cloud-init parameters: ipconfig0 and sshkeys.

In HCL2 syntax variable declarations contain various arguments as can be viewed in the documentation. This specific question is around the default argument, which supplies a value for the default value for a variable in a Packer template. This value must not be a dynamic expression, interpolated, etc.

In this situation, the values are literal strings such as #{sshkey}#. This signifies that literal string will be passed as the default value for the variable, which almost certainly will cause a runtime error against the source and/or build API as the value is a placeholder. These values would all need to be replaced with Packer input variables or variables files during an execution. If you are using Azure DevOps, I would have to assume this occurs within the pipeline where the values or files are passed as arguments to packer build or other Packer commands.

These default argument values currently would cause runtime errors if a value is not input for each, because they falsely cause the variables to be interpreted as optional instead of required. Best practices would involve removing these default arguments such that Packer would error during compilation/pre-check instead of runtime if an input is missing. This would improve workflow efficiency and debugging.

The best guess as to why these values were stubbed in the default arguments is perhaps to avoid missing variable input errors during a packer validate as part of CI or personal workflows. However, if packer validate is executed on the project directory instead of on the specific template and *.pkrvars.hcl are present, or if a file is supplied as an argument for the variables file, Packer would then also successfully validate the template(s) without these side issues.

The appleboy/ssh-action is supposed to use as parameter a host to contact, using the ssh key or a login/passord.

But once the connection is established, the commands done depends on the remote environment: if that git repository on the SSH-accessed remote machine has an HTTPS URL, no SSH key will be used.

Even if it has an SSH key, that still would not use any key from the ssh-action plugin.
It would use the ~remoteUser/.ssh/id_rsa default key already stored on the remote server home account.
If that key is there, to allow access to the remote GitHub private repository, then you need to change its origin to use the SSH URL:

This loop...