"Transport endpoint is not connected" on shell

[tbessie]

I posted this in another thread, but am starting a new one here...

I'm seeing "Transport endpoint is not connected" when I SSH to the shell server.

I called Sonic support, and they said they had no idea what to do in this situation, and that I should post here. Honestly, support should be able to handle this, it's not rocket science... :-/

I recently killed a bunch of old processes in my shell account, and apparently that killed support processes for the shell account itself.

Shouldn't these processes 1) not be killable by a user, and 2) auto-restart when killed?

- Tim

[scott]

I posted this in another thread, but am starting a new one here...

I'm seeing "Transport endpoint is not connected" when I SSH to the shell server.

I called Sonic support, and they said they had no idea what to do in this situation, and that I should post here. Honestly, support should be able to handle this, it's not rocket science... :-/

Unfortunately, tech support doesn't have access on the shell server to be able to unmount these stale mounts. For now, if you run into this situation, post here or email shellmaster@sonic.net.

I recently killed a bunch of old processes in my shell account, and apparently that killed support processes for the shell account itself.

Shouldn't these processes 1) not be killable by a user, and 2) auto-restart when killed?

Because of the way sshfs works, the ssh processes are owned by the user for whom they run. As you found out, that means you can kill them.

One of my next projects to do on the shell server is to revamp the mount management. Meanwhile, I will add a check for this condition and have the system automatically fix it.

-Scott

[tbessie]

Alright, thanks for explaining! Just curious, but why isn't support equipped to handle these situations? Why split out support into what the folks on the main support lines can do, and what you (and others) can do?

- Tim