@Migueldeicaza if I actively exit/logoff, should the preserver not exit completely?

It isn't for me, and it's not being reused when a new connection is made, so if you disconnect cleanly and reconnect each time you end up spawning another preserver....

@Migueldeicaza I was worried perhaps it was some weird interaction with docker running on the first host, but it behaves the same on another very plain Deb11 server.

A third Deb11 server (this one is admittedly.... special), the shell never starts, I just get connection log that ends in "The initial connection to the preservator has now closed" and no launched background process. I can't see any evidence of the preservator attempting to "re-ssh" to localhost... how would I generate more debug?

@Migueldeicaza it's a Deb11 host like the other two.

The binary itself ran ok;
preservator bb60eb8d1c369f26cb7b1767b40ccf924d6ab977

I only said this one is "special" because it is a legacy OSE which was previously a 32-bit that I did an in-place switch to 64-bit a while ago. I straced the binary and it's pulling in x64 libc, though, so no, it's not related to that.

Then I figured it out, by trying to manually SSH to 127.0.0.1 - this sshd is only listening to specific IPs. 🤦‍♂️

@Migueldeicaza I don't know if it's worth having a fallback attempt to ssh to the same IP the LaTerminal client is connecting to rather than localhost to avoid non-standard ListenAddress configurations, but I guess there's also a non-zero risk that might just end up bouncing off firewall rules anyway.

(plus I guess it is maybe less useful where that IP is dynamic in some way, if it changes then the preservator session will be lost.. I guess if that fallback was used, print a warning?)