On Wed, May 18, 2016 at 4:50 PM, Felix Miata <mrmazda@xxxxxxxxxxxxx> wrote: > Chris Murphy composed on 2016-05-18 16:32 (UTC-0600): > >> Chris Murphy> wrote: > > >>> 90 second hang at reboot, lsof reports many user session processes >>> aren't quitting >>> https://bugzilla.redhat.com/show_bug.cgi?id=1337307 > > >>> There are a bunch of services running as user chris that do not quit >>> well after a log out. I think that's what's causing exactly 1m30s hang >>> when trying to reboot or shutdown. Is anyone else seeing this? Was it >>> a clean install or an upgrade? > > >>> My case is an upgrade with Gnome Software. It wasn't happening with >>> Fedora 23 before the upgrade. I haven't yet tried a clean install of >>> Fedora 24. I'd think if this were widespread there'd be complaints by >>> now. > > >> Well it happens with a clean installation of >> Fedora-Workstation-Live-x86_64-24-20160518.n.0.iso so I'm not sure >> what's up... but whatever the cause is, systemd is giving up on >> whatever won't quit on its own exactly 1m30s after I ask for a >> shutdown every time. The log is not exactly revealing even with >> systemd.log_level=debug enabled. > > > Does the delay disappear if you immediately precede shutdown order with > 'umount -a -t cifs -t nfs' or some variant or subset thereof? No. reboot -f does work instantly so it's definitely not a kernel problem. Basically something(s) isn't quitting in the user session and so systemd is waiting for the user session to die or whatever the term is, and then it'll continue with reboot. But that user session doesn't ever die, and therefore the 1m30 systemd timeout applies for the user session and is just kills the session and reboots. The only thing mounted is a local file system. It's super basic. -- Chris Murphy -- test mailing list test@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe: http://lists.fedoraproject.org/admin/lists/test@xxxxxxxxxxxxxxxxxxxxxxx
