Open MartinLoeper opened 1 year ago
Hello, thanks for posting those logs. I did some testing with socket activation enabled and while I was not able to get the exact same error message, I have experienced similar behavior where it takes a second connection attempt for the ssh session to start working. This is most definitely a bug we want to fix, since there is no reason why we should not be able to support systemd socket activation.
Describe the bug
I am building a yocto distribution using AWS Greengrass V2 and AWS Secure-Tunneling by using the latest recipes from meta-aws. My tunnel is configured to connect to an SSH daemon on port 22. The daemon is started using systemd's sshd.socket activation (OpenSSH Per-Connection Daemon).
I am consistently experiencing that the first connection fails with the following log messages:
Right afterwards the destination is disconnected. The SSH daemon logs the following:
It looks like for the first connection attempt no per-connection daemon process exists and the localproxy crashes whereas for the second attempt the sshd process is started timely and localproxy is able to establish the connection.
Has anybody an idea what might be going on here? Is localproxy supposed to work with systemd socket activation?
Logs
see above
Environment (please complete the following information):