[Bug 857031] New: systemd: if NFS mounts are present, system hangs on shutdown when using ifup due to interfaces stopped too early
https://bugzilla.novell.com/show_bug.cgi?id=857031 https://bugzilla.novell.com/show_bug.cgi?id=857031#c0 Summary: systemd: if NFS mounts are present, system hangs on shutdown when using ifup due to interfaces stopped too early Classification: openSUSE Product: openSUSE 13.1 Version: Final Platform: Other OS/Version: openSUSE 13.1 Status: NEW Severity: Major Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: arvidjaar@gmail.com QAContact: qa-bugs@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0 With ifup each interface is started as separate service. These services do not have any ordering dependencies which means that on shutdown they are stopped as soon as possible. If NFS mounts are present, system will hang on shutdown for prolonged time (it will eventually reboot, not sure whether this is done by kernel or systemd). The following is suggested fix. All network mounts automatically get dependency on network-online.target. So let's order network@if.service before network-online.target on startup, which ensures remote filesystems are unmounted before interfaces are stopped: [Unit] .. Before=network-online.target Wants=network-online.target Wants here is required for manual NFS mounts, otherwise network-online.target is not started at all. I have tested it successfully, including busy mount points. If we are OK with this approach, SR will follow. Reproducible: Always Steps to Reproduce: 1. 2. 3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c
Andrey Borzenkov
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c1
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c2
Neil Brown
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c3
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c4
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c5
--- Comment #5 from Andrey Borzenkov
Now we have no network@.service anymore but several wicked serivces
wicked.service
If "wicked ifup ..." is synchronous, this should be already OK - wicked.service is ordered before network.target and pulls it. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c6
--- Comment #6 from Marius Tomaschewski
Hmm ... only the switch sysconfig-network over to wicked bothers me as the old
network@.service
was part of sysconfig-network. Now we have no network@.service anymore but several wicked serivces
wicked.service
This is the "oneshot" starter/stopper, which provides "network.service" similar to network.service on 13.1 [which triggered network@ starts]. It pushes config to the daemons, waits a while for them to start and the rest is done/continues inside of them.
wickedd-auto4.service wickedd-dhcp4.service wickedd-dhcp6.service wickedd-nanny.service wickedd.service
All wickedd* service can be considered as one. "systemctl restart wickedd" restarts all of them. Stopping them does not stop the network.service (wicked.service) and does not trigger network shutdown.
Also I'd like to submit a fixed sysconfig-network for 13.1
Sure, when we have something that works, we should submit it. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c7
--- Comment #7 from Marius Tomaschewski
(In reply to comment #3)
Now we have no network@.service anymore but several wicked serivces
wicked.service
If "wicked ifup ..." is synchronous, this should be already OK - wicked.service is ordered before network.target and pulls it.
It waits until the interfaces are up and running [when possible inside of a timeout]. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c8
Marius Tomaschewski
I have tested it successfully, including busy mount points. If we are OK with this approach, SR will follow.
Can you use "git format-patch" for the the patches on top of: https://github.com/openSUSE/sysconfig/tree/opensuse-13.1[-test] -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c10
Andrey Borzenkov
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c12
--- Comment #12 from Bernhard Wiedemann
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c13
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c14
--- Comment #14 from Swamp Workflow Management
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c15
Ruediger Meier
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c
Bernhard Wiedemann
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c16
Rodney Baker
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c17
--- Comment #17 from Rodney Baker
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c18
--- Comment #18 from Andrey Borzenkov
https://bugzilla.novell.com/show_bug.cgi?id=857031
https://bugzilla.novell.com/show_bug.cgi?id=857031#c19
--- Comment #19 from Rodney Baker
participants (1)
-
bugzilla_noreply@novell.com