What | Removed | Added |
---|---|---|
Assignee | mt@suse.com | mfilka@suse.com |
(In reply to Ralf Friedl from comment #0) > Actually in this case it is not even necessary to reconfigure the network ack. I think yast2 is simply unable to detect what it effectively changed. > But anyway, it should be possible to call "rcnetwork restart" without > causing that terminal to hang. It is "magic" why systemd sometimes results in a systemd-tty-ask wanting something. The network scripts themself never request any passwords which would cause to ask user for something. A "systemd-tty-ask-password-agent --list" could perhaps show what really needs a password from systemd. It could be something "foreign" started via if-up.d hook scripts or even more probably some service doing it gets restarted because of dependencies. The --ignore-dependencies and/or --no-ask-password options could help here [can be also passed to network script via SYSTEMCTL_OPTIONS variable]... Similar to DHCLIENT_MODIFY_SMB_CONF=yes in network/dhcp as a known source to cause dependency loops (systemd blocks an action to a service because an another in dependency chain is currently running an action) -- the if-up.d script handling it is installed by samba-client... > It seems the second attempt at rcnetwork restart managed to unconfigure > the network, not I don't even get responses to pings and therefor can't > open another ssh connection to the server. This is probably the garbage collector in systemd: kills services when they do not response for a while -- here, caused by the systemd-tty-ask.