http://bugzilla.suse.com/show_bug.cgi?id=1117217 http://bugzilla.suse.com/show_bug.cgi?id=1117217#c8 Thomas Blume <thomas.blume@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(obermiller@schwei | |zerbart.de) --- Comment #8 from Thomas Blume <thomas.blume@suse.com> --- (In reply to Walter Obermiller from comment #7)

The /var/run symlink to /run is there.

rpcbind _is_ running when I log in after boot (see comment4).

Sorry, I've overlooked that. The logs in comment#4 shows that you have enabled rpcbind.service: --> ● rpcbind.service - RPC Bind Loaded: loaded (/usr/lib/systemd/system/rpcbind.service; enabled; vendor preset: disabled) --< Normally that shouldn't be necessary since rpcbind is using socket activation. The only thing that should be enabled is rpcbind.socket. Please check via: systemctl status rpcbind.socket that rpcbind.socket is enabled and disable rpcbind.service.

ypbind is NOT running at that time (although enabled, see yptest output and systemctl status in comment 4)!

This puzzling observation caused me to think that there might be a timing/systemd dependency issue with the system (freshly installed from bare metal).

If you need further information, let me know.

The above might be the reason of the initial failure of rpcbind. When ypbind.service is starting up at that time, it fails, because rpcbind is down. Using socket activation will start up rpcbind.service at socket access, e.g. just when ypbind.service needs it. That should prevent an ypbind.service failure. Can you please check whether using rpcbind socket activation will fix this? If not, please provide the output of: journalctl -axb from your machine. -- You are receiving this mail because: You are on the CC list for the bug.