Reducing latency before retrieving a local terminal when an SSH session has expired

When working on a remote server, I open an SSH connection in a terminal ssh remote_username@remote_host.

For some reason, if I left this connection idle for too long, the terminal becomes unresponsive, e.g. whatever key I hit, absolutely nothing happens.

But after a certain amount of time, I'm given back the hand on the underlying local terminal and I can work again:

remote_user@remote_server:~$ # <--- whatever I type, nothing happens
# many minutes later, this is printed:
remote_user@remote_server:~$ client_loop: send disconnect: Broken pipe
local_user@local_machine:~ # and I'm given back the hand on my local terminal

This amount of time is quite long (many minutes).

Hence my question: is there a way to parametrize this delay on Ubuntu 22.04.1 LTS?

Versioning

• ssh: OpenSSH_8.9p1 Ubuntu-3ubuntu0.1, OpenSSL 3.0.2 15 Mar 2022
• bash: GNU bash, version 5.1.16(1)-release (x86_64-pc-linux-gnu)

You might need to set the option ServerAliveInterval ... From man ssh_config:

ServerAliveInterval

Sets a timeout interval in seconds after which if no data has been received from the server, ssh(1) will send a message through the encrypted channel to request a response from the server. The default is 0, indicating that these messages will not be sent to the server.

It takes a value as a number of seconds ... e.g. like so:

ServerAliveInterval 120

You can add that in a new line to ssh config either for your user in the file ~/.ssh/config(It's not there by default so you might need to create this file) or globally in the file /etc/ssh/ssh_config ... Notice that you should add it under a Host keyword ... e.g. in the /etc/ssh/ssh_config there is a keyword Host * which means apply the below arguments to all remote hosts ... So you would add it to the existing arguments e.g. like so:

Host *
    SendEnv LANG LC_*
    HashKnownHosts yes
    GSSAPIAuthentication yes
    ServerAliveInterval 120

Alternatively, as stated in man ssh:

-o option

Can be used to give options in the format used in the configuration file.

You can set that option directly in the command itself ... e.g. like so:

ssh -o ServerAliveInterval=120 user@host

I see this sometimes with my home connection to a remote server. One workaround is to leave something running on the remote end which produces output, so traffic is constantly flowing, but not to much that it's a bandwidth hog. What I use is Byobu

While running, Byobu has a clock on screen (along with other things like disk space, and other useful data) in the status bar. Having that on screen typically stops the connection timing out.