[Bug 730193] New: shutdown/poweroff/reboot hangs when using sysvinit
https://bugzilla.novell.com/show_bug.cgi?id=730193 https://bugzilla.novell.com/show_bug.cgi?id=730193#c0 Summary: shutdown/poweroff/reboot hangs when using sysvinit Classification: openSUSE Product: openSUSE 12.1 Version: Final Platform: VMWare OS/Version: Other Status: NEW Severity: Normal Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: d.a.van.delft@gmail.com QAContact: qa@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; nl; rv:1.9.2.23) Gecko/20110920 SUSE/3.6.23-0.2.1 Firefox/3.6.23 It works OK when in systemd mode. When it hangs, switching between VT's still work, but they are unusable: cannot log in, or see what's going on. Reproducible: Always Steps to Reproduce: 1. install sysvinit-init 2. for good measure, reboot at least once to have it run on sysvinit 3. as root either shutdown, reboot of poweroff Actual Results: Starts shutdown sequence. Usually hangs after console message: Shutting down service (localfs) network . . . . . . done But sometimes others as well. When it hangs, the CPU usage goes to 100%, as can be seen on the vmware host. Expected Results: an actual shutdown/poweroff or reboot I can only reproduce this on two VMWare vm's, running on vmware server 2 and VSphere. A hardware computer works OK. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c
zj jia
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c1
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c2
Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c3
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c4
Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c5
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c6
Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c7
--- Comment #7 from Danny van Delft
Hmmm ... or the redirect to /dev/console results in invisible output. Now try to add
set -x
after the line
set +e
in /etc/init.d/halt ... let us see what happens upto the very last message
OK, did that with kernel-desktop. Last lines of output: Running /etc/init.d/halt.local + /bin/sh /etc/init.d/halt.local after which nothing more appears. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c8
--- Comment #8 from Danny van Delft
Hmmm ... or the redirect to /dev/console results in invisible output. Now try to add
set -x
after the line
set +e
in /etc/init.d/halt ... let us see what happens upto the very last message
Done the same with kernel-default, now different output appears. Looks like another codepath, which succeeds, is taken in comparison to kernel-desktop. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c9
--- Comment #9 from Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c10
--- Comment #10 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c11
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c12
--- Comment #12 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c13
--- Comment #13 from Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c14
--- Comment #14 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c15
Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c16
--- Comment #16 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c17
--- Comment #17 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c18
--- Comment #18 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c19
--- Comment #19 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c20
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c22
--- Comment #22 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c23
Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c24
--- Comment #24 from Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c25
--- Comment #25 from Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c26
Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c27
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c28
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c29
Andreas Jaeger
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c30
--- Comment #30 from Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c31
Michael Matz
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c32
--- Comment #32 from Lars Müller
(In reply to comment #26)
Lars? Why do you think this?
As it works with systemd. Therefore I strongly believe this is something caused by sysvinit. Well, you raised the right question. This is nothing I can prove. All I did is guessing from the symptoms. Therefore I think the best is we close this issue and forget about it? No, we have to track it down to the roots. And here the trouble starts. The system in question is in a data center. Where I don't have access at any time. Therefor we need to script something which does the same as described in comment 15. Unfortunately I don't have the time to do this at the moment.
A syou can see from my comment #24 this is a kernel bug. Or why does this never happen with a none VM system?
This is not a virtual machine. Or does your VM reference to something else?
I've several systems around here using SysVinit on 12.1 and those systems do *not* hang.
I also tested this on three other native 12.1 systems. Both are x86_64 too And on both I'm also not able to cause the same issue. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c33
Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c34
Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c35
--- Comment #35 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c36
Lars Müller
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c37
--- Comment #37 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c38
--- Comment #38 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c39
Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c40
--- Comment #40 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c41
--- Comment #41 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c42
--- Comment #42 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c43
--- Comment #43 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c44
--- Comment #44 from Matthias Pfafferodt
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c45
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c46
Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c47
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c48
Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c49
--- Comment #49 from Jeff Mahoney
(In reply to comment #39)
For NFS shares as well as for any other network based shared filessystem:
If the server of such an share is not reachable then any system call like e.g. stat(2) will run into a deadlock. I've added a lot of workarounds in the last few years to avoid system call calls like stat(2) but even by using readlink(2): if the kernel does not have any information cached it will ask the server of the share and this also lead to a deadlock.
To be fair, this isn't an actual deadlock. It's a network dependency that will be met when the server comes back. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c50
--- Comment #50 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c51
--- Comment #51 from Matthias Pfafferodt
(In reply to comment #39)
For NFS shares as well as for any other network based shared filessystem:
If the server of such an share is not reachable then any system call like e.g. stat(2) will run into a deadlock. I've added a lot of workarounds in the last few years to avoid system call calls like stat(2) but even by using readlink(2): if the kernel does not have any information cached it will ask the server of the share and this also lead to a deadlock.
This has nothing todo with sysvinit its self. It may depend on the execution order at shuthdown (that is that all network connections should go down *after* the the network based share is unmounted). That is you should check the dependcy order of the boot scripts and enforece that any process on or with the network based shares are stopped *before* the network for those shares will switched off.
This means that using nfs mounted home (nis and autofs) I will have to ensure that all processes accessing the nfs share are killed before doing a shutdown? The server was online at the time of the shutdown so this is not the reason. Thus, the order of the shutdown scripts should be changed? I have to say, that I do not understand how the boot scripts network-remotefs and network interact.
IMHO the initial report does not mention a network based share. That is your report does not belong to this bug, does it?
Yes, my last report does depend on using an nfs share and also in my first post all linux instances used to have an nfs share mounted. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c52
Dr. Werner Fink
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c53
--- Comment #53 from Marius Tomaschewski
This means that using nfs mounted home (nis and autofs) I will have to ensure that all processes accessing the nfs share are killed before doing a shutdown?
Not exactly you, but the nfs/autofs/remomte-fs-depending... service scripts. So the question is, what is still running at the time nfs and network go down. -- 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=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c54
--- Comment #54 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c55
--- Comment #55 from Ludwig Nussel
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c56
--- Comment #56 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c57
--- Comment #57 from Ludwig Nussel
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c58
--- Comment #58 from Danny van Delft
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c59
Bernhard Wiedemann
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c60
--- Comment #60 from Benjamin Brunner
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c61
--- Comment #61 from Bernhard Wiedemann
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c62
--- Comment #62 from Swamp Workflow Management
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c63
Swamp Workflow Management
https://bugzilla.novell.com/show_bug.cgi?id=730193
https://bugzilla.novell.com/show_bug.cgi?id=730193#c64
Ludwig Nussel
participants (1)
-
bugzilla_noreply@novell.com