Red Hat Bugzilla – Bug 1270369
On shutdown, systemd tries to unmount network shares after NetworkManager is stopped, resulting in a long delay on wireless
Last modified: 2015-10-12 03:36:44 EDT
On my F22 laptop, I have a couple of shares in /etc/fstab :
//nas/Media /share/data cifs noauto,nosuid,soft,guest,uid=99,gid=99,file_mode=0777,dir_mode=0777,users,comment=systemd.automount 0 0
ircproxy:/ /share/irclogs nfs noauto,user,comment=systemd.automount 0 0
it seems that on shutdown, systemd attempts to stop these after NetworkManager has already stopped. Obviously at that point it's not going to be able to unmount them cleanly.
Attaching a log: you can see that shutdown is triggered at 10:28:57, NetworkManager is taken down in the same second, and errors from CIFS and NFS start showing up around 10:30:11:
Oct 09 10:30:11 xps13.happyassassin.net kernel: CIFS VFS: Server nas has not responded in 120 seconds. Reconnecting...
by 10:30:47 it's given up on the CIFS mount:
Oct 09 10:30:47 xps13.happyassassin.net systemd: Unmounted /share/data.
but the NFS mount holds it up much much longer. At 10:31 there's this message:
Oct 09 10:31:58 xps13.happyassassin.net systemd: share-irclogs.mount unmounting timed out. Stopping.
but in fact it just keeps hitting NFS timeouts right up until the reboot.target job times out, half an hour after it started:
Oct 09 10:58:57 xps13.happyassassin.net systemd: Job reboot.target/start timed out.
whereupon the system finally shuts down.
Created attachment 1081404 [details]
journal from an affected boot
So I think the connection being wireless is significant here. On a wired connection, even after NetworkManager.service is stopped, the connection remains active. On a wireless connection this is not the case.
*** This bug has been marked as a duplicate of bug 1214466 ***