[Bug 963616] New: Boot fails to wait for a needed fstab NFS mount, fires a 'wicked nanny' error ("Failed to start Alias for NFS client.". systemd unit dependency fix included.
http://bugzilla.opensuse.org/show_bug.cgi?id=963616 Bug ID: 963616 Summary: Boot fails to wait for a needed fstab NFS mount, fires a 'wicked nanny' error ("Failed to start Alias for NFS client.". systemd unit dependency fix included. Classification: openSUSE Product: openSUSE Distribution Version: Leap 42.1 Hardware: x86-64 OS: openSUSE 42.1 Status: NEW Severity: Major Priority: P5 - None Component: Network Assignee: neilb@suse.com Reporter: suse.dev@fea.st QA Contact: qa-bugs@suse.de CC: wicked-maintainers@suse.de Found By: --- Blocker: --- I'm booting a VM guest that needs to bring up an NFS mount. Right now on boot, something's out of order and it fails ... [ OK ] Started wicked network management service daemon. [ 23.805688] systemd[1]: Started wicked network management service daemon. [ 23.834962] systemd[1]: Starting wicked network nanny service... Starting wicked network nanny service... [FAILED] Failed to start Alias for NFS client. See "systemctl status nfs.service" for details. [ 23.876106] systemd[1]: Failed to start Alias for NFS client. [ OK ] Started System Logging Service. [ OK ] Started wicked network nanny service. ... Checking nfs.service - Alias for NFS client Loaded: loaded (/usr/lib/systemd/system/nfs.service; enabled) Drop-In: /run/systemd/generator/nfs.service.d └─50-insserv.conf-$remote_fs.conf Active: failed (Result: exit-code) since Tue 2016-01-26 07:16:26 PST; 52s ago Process: 738 ExecStop=/usr/bin/umount -aft nfs,nfs4 (code=exited, status=0/SUCCESS) Process: 727 ExecStartPost=/usr/bin/mount -at nfs,nfs4 (code=exited, status=32) Process: 705 ExecStart=/bin/true (code=exited, status=0/SUCCESS) Main PID: 705 (code=exited, status=0/SUCCESS) Jan 26 07:16:26 linux-u398 mount[727]: mount.nfs4: Failed to resolve server nfs01.example.com: Name or service not known That server's the mount defined in the guest's /etc/fstab cat /etc/fstab ... nfs01.example.com:/SEC/clients/client06 /sec/client06 nfs4 _netdev,auto,ro,soft,tcp,async,noexec,nosuid,bg,rsize=32768,wsize=32768 0 0 IIUC, fstab entries are translated by systemd-fstab-generator into native systemd units durint early boot, and it's systemd that controls the mounts & their dependencies. But, here, wicked's also involved. After talking in #systemd, this fix works edit /etc/systemd/system/nfs.service.d/deps.conf [Unit] Requires=network-online.target After=network-online.target After=remote-fs.target reboot Now, the mount's correctly mounted on boot, with no more error in logs. I don't know if those are the best/only dependencies that should be there. I only know that it manages to fix this problem. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=963616
http://bugzilla.opensuse.org/show_bug.cgi?id=963616#c1
Neil Brown
http://bugzilla.opensuse.org/show_bug.cgi?id=963616
http://bugzilla.opensuse.org/show_bug.cgi?id=963616#c2
sd fs
The fix doesn't seem right to me. Making nfs.service be "After=remote-fs.target" seems very odd as nfs is needed for remote-fs...
Yeah, NFS deps have my head spinning a bit. Like I said -- it 'just works' right now. If it's the best or not, or breaks something else, just don't know yet.
Requiring network-online makes a bit more sense, but it isn't really the nfs.service but rather the mounts that should require that.
Where is your name service happening? I doubt that you are running named on the VM but could you please confirm. There was a problem where named wasn't started until after remote-fs and that caused problems much like this. I think that was fixed but maybe the fix isn't in Leap yet.
I'll try to duplicate...
NS here is on the LAN, same subnet, but on different hardware. I.e., not on the DomU and not on the Dom0. Another box entirely. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=963616
http://bugzilla.opensuse.org/show_bug.cgi?id=963616#c3
Neil Brown
http://bugzilla.opensuse.org/show_bug.cgi?id=963616
http://bugzilla.opensuse.org/show_bug.cgi?id=963616#c4
sd fs
... and I failed to duplicate it.
Yeah, bot surprised. Guessing timings/timeouts matter and something's near an edge.
Are you able to perform a test for me? Remove the "deps.conf" and create instead a "post.conf" file in the same directory containing
[Service] ExecStartPost=
that will disable the "/usr/bin/mount -at nfs,nfs4" line that doesn't seem to be necessary and may be causing a problem. Then reboot.
Making the change rm /etc/systemd/system/nfs.service.d/deps.conf edit /etc/systemd/system/nfs.service.d/post.conf [Service] ExecStartPost= reboot guest at host console ... [ OK ] Started Name Service Cache Daemon. [ 23.963669] systemd[1]: Started Name Service Cache Daemon. [ OK ] Started Alias for NFS client. [ 23.997343] systemd[1]: Started Alias for NFS client. [ 24.027148] systemd[1]: Starting Remote File Systems (Pre). [ OK ] Reached target Remote File Systems (Pre). [ 24.059333] systemd[1]: Reached target Remote File Systems (Pre). ... [ OK ] Started OpenSSH Daemon. [ 40.888124] systemd[1]: Started OpenSSH Daemon. ... login at guest console systemctl status nfs.service nfs.service - Alias for NFS client Loaded: loaded (/usr/lib/systemd/system/nfs.service; enabled) Drop-In: /run/systemd/generator/nfs.service.d └─50-insserv.conf-$remote_fs.conf /etc/systemd/system/nfs.service.d └─requires.conf Active: active (exited) since Wed 2016-01-27 06:18:41 PST; 4min 45s ago Process: 732 ExecStart=/bin/true (code=exited, status=0/SUCCESS) Main PID: 732 (code=exited, status=0/SUCCESS) CGroup: /system.slice/nfs.service So, seems to work -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=963616
http://bugzilla.opensuse.org/show_bug.cgi?id=963616#c5
Neil Brown
participants (1)
-
bugzilla_noreply＠novell.com