Today in "Fun with your homelab":
One of my openSUSE MicroOS machines was under heavy load for a week and I could not find out what was wrong. I noticed that SSH logins were no longer possible, the login hung after the "Last login ..." line. Non-interactive logins worked, though.
Today I found out that suddenly this machine requires "ssh -t" or the RequestTTY option to get a working SSH connection. Huh?
All other MicroOS machines are of course unaffected by this (SSH from the same laptop to those machines works without "-t"), otherwise life would not be interesting...
Debugging this and then checking where the high load comes from...
#Linux #SSH #openSUSE #MicroOS #homelab #HellYeah #AdminLife

Attached: 1 image Dons ja tenim migrat el servidor #Mastodon amb #Quadlets a #MicroOS a #OpenSUSE

I've managed to track this to [email protected] failing to start for some reason. Upon ssh, it starts properly.
Is this expected #systemd behavior? This sure doesn't look like it. I'm suspecting this is a #MicroOS issue.
In recent times my Raspberry Pi server with #MicroOS has been behaving weirdly.
My server sometimes goes down and it seems it's because of an automated system restart according to the logs (when I have transactional-update disabled, so that shouldn't be happening).
When that happens, it takes a bit for me to ssh to it. It seems like my #podman #quadlets are not obeying #systemd enable-linger and starting on boot with WantedBy=default.target , because they only start after I ssh to it.