NMI watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [gdbus:2878]

[Arlie Stephens]

This is a bug report about errors during the upgrade process. I don't know whether or not my samba installation was working immediately before the upgrade; the windows box that was intended to be a client died and has not been replaced. Don't believe my answers to the forced questions.

***

My ubuntu system locks up regularly. Bugs I've reported about this get closed for inattention, after being ignored, and queries about how to do basic debugging get the same treatment, so I just power cycle whenever it happens again.

This time the lock up happened while the GUI upgrade-everything-to-latest tool was running.

When I power cycled the box to recover, the system was unusable - flashing "system problem detected" popups, none staying stable long enough to respond.

ssh + "sudo apt-get update" + "sudo apt-get upgrade" pointed me to "sudo dpkg --configure -a"

That command reported numerous errors - I don't know which of them are "normal" for this system, but hidden by the GUI, and which are new.

But it got the system semi-usable from the GUI, so now I'm clicking on "yes, report it" for the many pop-ups, of which this is one.

FWIW, samba is one of the ones "dpkg --reconfigure -a" complained about.

ProblemType: Package
DistroRelease: Ubuntu 16.04
Package: samba 2:4.3.11+dfsg-0ubuntu0.16.04.10
ProcVersionSignature: Ubuntu 4.4.0-93.116-generic 4.4.79
Uname: Linux 4.4.0-93-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.20.1-0ubuntu2.10
Architecture: amd64
Date: Sun Nov 12 11:07:42 2017
ErrorMessage: subprocess new pre-removal script returned error exit status 1
InstallationDate: Installed on 2012-04-26 (2026 days ago)
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Release amd64 (20120425)
NmbdLog:

OtherFailedConnect: Yes
RelatedPackageVersions:
 dpkg 1.18.4ubuntu1.2
 apt 1.2.24
SambaServerRegression: Yes
SmbConfIncluded: Yes
SmbLog:

SourcePackage: samba
Title: package samba 2:4.3.11+dfsg-0ubuntu0.16.04.10 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1
UpgradeStatus: Upgraded to xenial on 2017-03-31 (226 days ago)

[Arlie Stephens (arlie)]

The bug detection and reporting system wants to report this bug twice ;-) It's the same incident, but perhaps the actual error managed to happen twice somehow.

[Andreas Hasenack (ahasenack)]

Thanks for filing this bug in Ubuntu.

Sorry for the troubles you are having, but indeed your logs show a lot of "soft lockups" in the kernel, and all seem to be in the nvidia module.

I'm going to change this bug as affecting nvidia, since I don't believe it's samba specific.

[Andreas Hasenack (ahasenack)]

Sorry, it's actually gdbus that is soft locking

[Arlie Stephens (arlie)]

Happened again. Auto-bug reporting didn't recognize this report as similar, so created bug 1734405

Note - compiz was in Z state (per top), consuming 100% of a cpu (per top) throughout the incident. I was able to login via ssh and look at things.

I don't believe that a stuck compiz process could be caused by soft lockups. So the compiz process *may* be closer to the root cause?

[Andreas Hasenack (ahasenack)]

Or nvidia

[Andreas Hasenack (ahasenack)]

I added an nvidia task to the bug so its maintainer can take a look to see if it's related or not.