Figure I'd post this fix for anyone that has a similar issue migrating from 5.4 (Stretch) to 6 (Buster).
I have a standalone PVE node, originally installed on Debian Stretch a few years ago with no major mods outside of a default installation. Ran pve5to6 with no flags, then the update.
After the update, networking would fail exactly 22 minutes (give or take 5 seconds) after boot - no ssh, webgui, etc. Everything looked and seemed totally normal up until that mark. Checked crons, and a newly installed UPS. Ended up being a new power feature in Debian 10 Buster...
Apparently in Buster, the default power settings suspends after 20 minutes for power savings. Stretch did not have this feature, so your default config into Buster will likely be the same aka default suspending after 20 mins. When the system suspends it suspends NetworkManager (like windows workstations
Following the advice here worked for me since I need network access 24/7:
systemctl mask sleep.target suspend.target hibernate.target hybrid-sleep.target
I have a standalone PVE node, originally installed on Debian Stretch a few years ago with no major mods outside of a default installation. Ran pve5to6 with no flags, then the update.
After the update, networking would fail exactly 22 minutes (give or take 5 seconds) after boot - no ssh, webgui, etc. Everything looked and seemed totally normal up until that mark. Checked crons, and a newly installed UPS. Ended up being a new power feature in Debian 10 Buster...
Apparently in Buster, the default power settings suspends after 20 minutes for power savings. Stretch did not have this feature, so your default config into Buster will likely be the same aka default suspending after 20 mins. When the system suspends it suspends NetworkManager (like windows workstations
Following the advice here worked for me since I need network access 24/7:
systemctl mask sleep.target suspend.target hibernate.target hybrid-sleep.target
Last edited: