ssh_known_hosts | Development repository for Chef Cookbook | Configuration Management library

in case I stop the sshd on Win10 and start it not as a service, just with c:>sshd.exe

• In one case, sshd is run as a service, possibly with the LocalSystem account
• In the other, from comand-line, you are running sshd as you (your Windows account).

The environment would not be the same, especially for the "authorized_keys" file.
As mentioned in the documentation:

In Windows, sshd reads configuration data from %programdata%\ssh\sshd_config by default, or a different configuration file may be specified by launching sshd.exe with the -f parameter.
If the file is absent, sshd generates one with the default configuration when the service is started.

Activate the logs when running the ssh daemon as a Windows service, as seen here:

Open an admin powershell prompt and run:

If the private key format differs, that means, as I mentioned here that:

• one platform is using openssh prior to 7.8, with an old PEM 64-chars per line format.
• one is using a more recent OpenSSH format, 70-chars per line.

You can force a recent openSSH to generate the old format with:

I realized that if I removed the target definitions altogether and let the arm64 docker environment build the rust app it works as expected. Thanks to @user1937198, I found that using "aarch64-unknown-linux-musl" as the target allowed the static build of the arm rust in the container. The working docker file is below

This is definitely an SSH problem. The SSH(1) man page states:

ssh automatically maintains and checks a database containing identification for all hosts it has ever been used with. Host keys are stored in ~/.ssh/known_hosts in the user's home directory. Additionally, the file /etc/ssh/ssh_known_hosts is automatically checked for known hosts.

Your issue might be that the un-elevated user cannot read the /etc/ssh/ssh_known_hosts file.

There is a caveat to ssh where if you run something like sudo ssh-keygen -R domain.com, it can modify your existing /etc/ssh/ssh_known_hosts/ file to be only readable by root. (-rw------- root root). You might want to use chown or chmod to change permissions of this file to make sure it is readable without root privileges.

git@github.com:organization/my-package.git is a valid SSH URL. ssh://git@github.com:organization/my-package.git is not. ssh://git@github.com/organization/my-package.git would be.

As in here, you can add GIT_SSH_COMMAND='ssh -v' pip install ... to see exactly what is going on.

You might need:

I didn't manage to add a connection through visual studio IDE so I used the ConnectionManager.exe tool . at first I got an error because the date was in Arabic but after I changed the date the connection was added successfully .

I'm not sure about sshpass, since I usually use public/private keys. Here's an example of a job I would setup to run SCP/SSH commands on remote servers:

You should verify your host keys by running ssh-keyscan on the private server and copying the results to a GitLab variable called SSH_KNOWN_HOSTS. In your pipeline you need to copy that to the ~/.ssh/known_hosts file in the pipeline environment. Instructions are here: https://docs.gitlab.com/ee/ci/ssh_keys/#verifying-the-ssh-host-keys.

On a side note, you should consider creating a secure directory to store your config file in. See: https://docs.ansible.com/ansible/latest/reference_appendices/config.html

Okay well I feel stupid and solved it. I kept getting the permission denied because of the following issue: