First things first, I'm running a r14322 trunk server in Fedora 24, with the
$XDG_RUNTIME_DIR and the
systemd-run functionality enabled - so this should work fine, but when I log out of the SSH session, my server gets killed. After re-logging in via SSH, if I do an
xpra list, it finds an unknown session and then cleans it up. Disabling
--systemd-run=no makes the problem go away.
Found the following xpra sessions: /run/user/1001/xpra: UNKNOWN session at :13 /home/max/.xpra: UNKNOWN session at :13 Re-probing unknown sessions in: /run/user/1001/xpra, /home/max/.xpra UNKNOWN session at :13 (cleaned up) UNKNOWN session at :13 (cleaned up)
And, attempting to start a new server fails with the following message:
using systemd-run to wrap 'start' server command 'systemd-run' '--scope' '--user' '/usr/bin/xpra' 'start' ':13' '--bind-tcp=0.0.0.0:2200' '--start-child=xterm' '--daemon=no' Failed to create bus connection: No such file or directory
Note: It's failing to start the server whatsoever, so
-d flags do nothing.
The only way to be able to start a new server is to reboot the machine completely, or start one with
So, is there a setting somewhere that isn't set properly? I thought the point of using
systemd-run was such that we can run servers without having to keep an SSH session going. Or, is this not the point of
systemd-run? And, I'm using it incorrectly, or with the wrong expectations.
A side-effect of this is that if the network connection of the machine that started the session (like a laptop on spotty wi-fi) goes down, it takes the whole server with it, which kinda negates the point of daemonizing.
The correct ticket for reporting this issue is #1105.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/1348