Bug 30769

Was able to replicate with the LXQt desktop, so this is not an Xfce issue. 

Please note that I installed an Nvidia GT-710 video card in the open PCI-E x1 slot (the PC came with two x1 slots) and Linux would not boot, apparent IRQ conflict.


 ~]$ inxi -F
System:
  Host: upstairs Kernel: 5.15.58-desktop-2.mga8 x86_64 bits: 64 
  Desktop: Xfce 4.16.0 Distro: Mageia 8 mga8 
Machine:
  Type: Desktop System: Hewlett-Packard product: CQ5826 v: N/A 
  serial: <superuser required> 
  Mobo: FOXCONN model: 2AB7 v: 1.00 serial: <superuser required> 
  BIOS: American Megatrends v: 6.09 date: 05/04/2011 
CPU:
  Info: Dual Core model: AMD Athlon II X2 260 bits: 64 type: MCP 
  L2 cache: 2 MiB 
  Speed: 800 MHz min/max: 800/3200 MHz Core speeds (MHz): 1: 800 2: 800 
Graphics:
  Device-1: AMD RS780L [Radeon 3000] driver: radeon v: kernel 
  Display: x11 server: Mageia X.org 1.20.14 driver: ati,radeon,v4l 
  resolution: 1680x1050~60Hz 
  OpenGL: 
  renderer: AMD RS780 (DRM 2.50.0 / 5.15.58-desktop-2.mga8 LLVM 11.0.1) 
  v: 3.3 Mesa 21.3.8 
Audio:
  Device-1: AMD SBx00 Azalia driver: snd_hda_intel 
  Sound Server: ALSA v: k5.15.58-desktop-2.mga8 
Network:
  Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet 
  driver: r8169 
  IF: enp2s0 state: up speed: 1000 Mbps duplex: full mac: <removed> 
Drives:
  Local Storage: total: 465.76 GiB used: 106.54 GiB (22.9%) 
  ID-1: /dev/sda vendor: Western Digital model: WD5000AAKX-603CA0 
  size: 465.76 GiB 
Partition:
  ID-1: / size: 49.2 GiB used: 7.5 GiB (15.2%) fs: ext4 dev: /dev/sda1 
  ID-2: /home size: 403.94 GiB used: 99.04 GiB (24.5%) fs: ext4 
  dev: /dev/sda6 
Swap:
  ID-1: swap-1 type: partition size: 4 GiB used: 0 KiB (0.0%) dev: /dev/sda5 
Sensors:
  System Temperatures: cpu: 32.1 C mobo: N/A 
  Fan Speeds (RPM): N/A 
Info:
  Processes: 179 Uptime: 4m Memory: 7.53 GiB used: 1.13 GiB (15.0%) 
  Shell: Bash inxi: 3.2.01

Source RPM: xfce4-panel-4.16.1-1.1.mga8.src.rpm => (none)
Summary: xfce4 - Mageia 8 freezes upon minimizing/maximizing window from bottom Xfce panel => Mageia 8 freezes upon minimizing/maximizing window from bottom Xfce panel

Additional info:

Mageia 8 only freezes like this when it's running Oracle VirtualBox Manager with one VirtualBox running.

I have yet to replicate this not running VirtualBox.

Summary: Mageia 8 freezes upon minimizing/maximizing window from bottom desktop panel => Mageia 8 freezes upon minimizing/maximizing window from bottom desktop panel while Oracle VirtualBox Manager is open with one VirtualBox running

The interface will also occasionally freeze while web browsing using Firefox. I would click a link on a web page, then the interface freezes again.

Summary: Mageia 8 freezes upon minimizing/maximizing window from bottom desktop panel while Oracle VirtualBox Manager is open with one VirtualBox running => Mageia 8 freezes upon minimizing/maximizing window from bottom desktop panel while Oracle VirtualBox Manager is open with one VirtualBox running or will freeze while web browsing with Firefox

New freeze:

Had Audacious playing music, but minimized. Launched VirtualBox Manager, selected a VirtualBox, Audacious window button disappeared from Xfce panel, mga8 froze again.

Is is completely frozen, i.e not reacting to
§ ctrl-alt-bkspc,bkspc
§ ctrl-alt-del,del
§ press power button for 2 seconds (usually >1s make system shut down, but >4s cuts power directly per mainboard hardware function)

---

Can you boot into memory test, and run memory test over night, using all CPU cores?

That found the problem for me a couple years ago - I had a core that during that test occasionally experienced "memory fault", but not the other cores...

CC: (none) => fri

Also check to see if alt+ctrl+f2 will get to a text login where you can
login as root and check the output of the dmesg command.

CC: (none) => davidwhodgins

(In reply to Morgan Leijström from comment #5)
> Is is completely frozen, i.e not reacting to
> § ctrl-alt-bkspc,bkspc
> § ctrl-alt-del,del
> § press power button for 2 seconds (usually >1s make system shut down, but
> >4s cuts power directly per mainboard hardware function)
> 
> ---
> 
> Can you boot into memory test, and run memory test over night, using all CPU
> cores?
> 
> That found the problem for me a couple years ago - I had a core that during
> that test occasionally experienced "memory fault", but not the other cores...

I have been pressing the power button to power down and get the system back.

I have since reinstalled x11-driver-video-ati, but it probably will not make a difference.

CTL-ALT-BKSP, CLT-ALT-DEL, nothing. PRT SCN works - launches screenshot utility, but its GUI non-responsive. Mouse works. I can click anything on the screen, with no response.

The RAM is fairly new (2020) and had been tested a while back, showing no errors. I don't think the test took more than a couple of hours and AFAIK, it was a complete test. I can certainly run it again. IIRC, it is DDR3 SDRAM, two sticks, each with 4 Gb on them (motherboard accepts no more than 8Gb).

(In reply to Dave Hodgins from comment #6)
> Also check to see if alt+ctrl+f2 will get to a text login where you can
> login as root and check the output of the dmesg command.

Did that yesterday. Was able to get to the terminal, dmesg didn't show any obvious errors, nor did Xorg.0.log.

(In reply to Edward from comment #7)
> (In reply to Morgan Leijström from comment #5)
> I have been pressing the power button to power down and get the system back.

The interesting thing i asked about is if the system powers down after a short press, or you need to hold it until power cuts.


> The RAM is fairly new (2020) and had been tested a while back, showing no
> errors.

I also had new ram, but the memory tester also detect other problems in the data paths, such as connectors, and in my case it found a bad CPU core.


Anyway, your comment 8, terminal working, is what we can proceed on.

(In reply to Morgan Leijström from comment #9)
> (In reply to Edward from comment #7)
> > (In reply to Morgan Leijström from comment #5)
> > I have been pressing the power button to power down and get the system back.
> 
> The interesting thing i asked about is if the system powers down after a
> short press, or you need to hold it until power cuts.
> 
> 
> > The RAM is fairly new (2020) and had been tested a while back, showing no
> > errors.
> 
> I also had new ram, but the memory tester also detect other problems in the
> data paths, such as connectors, and in my case it found a bad CPU core.

I have the memory testing now, completed 1 pass already, no errors.

I have been holding in the power button to power it down. On occasion, probably if I'm not holding it in long enough, it will reboot instead

(In reply to Morgan Leijström from comment #5)

> Can you boot into memory test, and run memory test over night, using all CPU
> cores?
> 
> That found the problem for me a couple years ago - I had a core that during
> that test occasionally experienced "memory fault", but not the other cores...

The system is BIOS, no UEFI. The last available memtest86 image for a BIOS system with Linux, is a release under version 4. I did not see anything on the screens where it allowed me to test specifying the number of cores, the default was 1. It went through two complete passes, each had 9 or 10 different tests, took over two hours total and it reported no errors.

From comment 0:
> if a window is repeatedly minimized and maximized
From comment 2:
> only freezes like this when it's running Oracle VirtualBox Manager
> with one VirtualBox running
It looks as if (the Firefox issue apart) you have to do specific things to create the fault, so it is not 'critical' in the sense of not allowing any vital but normal system use.

Severity: critical => major

VirtualBox 6.1.38 also freezing the Mageia 8 UI. Minimized the VBox window, ONCE, to access something on the host and that was enough to freeze the UI.

As to Firefox freezing the UI, separately, I believe is related to a local television channel web site, it has too many videos on it. I was able to replicate the freeze multiple times via the same web site with 6.1.36.

Would like to confirm this is expected, when removing via urpme --auto-orphans:


 ~]$ sudo urpme --auto-orphans
To satisfy dependencies, the following 4 packages will be removed (157MB):
  
(orphan packages)
  kernel-desktop-5.15.58-2.mga8-1-1.mga8.x86_64
  kernel-desktop-5.15.62-1.mga8-1-1.mga8.x86_64
  virtualbox-kernel-5.15.58-desktop-2.mga8-6.1.36-1.4.mga8.x86_64
  virtualbox-kernel-5.15.62-desktop-1.mga8-6.1.38-1.2.mga8.x86_64
Remove 4 packages? (y/N) y
removing kernel-desktop-5.15.58-2.mga8-1-1.mga8.x86_64 kernel-desktop-5.15.62-1.mga8-1-1.mga8.x86_64 virtualbox-kernel-5.15.58-desktop-2.mga8-6.1.36-1.4.mga8.x86_64 virtualbox-kernel-5.15.62-desktop-1.mga8-6.1.38-1.2.mga8.x86_64

-------- Uninstall Beginning --------
Module:  virtualbox
Version: 6.1.38-1.2.mga8
Kernel:  5.15.62-desktop-1.mga8 (x86_64)
-------------------------------------

Status: Before uninstall, this module version was ACTIVE on this kernel.

vboxdrv.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.62-desktop-1.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.

vboxnetflt.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.62-desktop-1.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.

vboxnetadp.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.62-desktop-1.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.
depmod.........

DKMS: uninstall Completed.
removing package virtualbox-kernel-5.15.62-desktop-1.mga8-6.1.38-1.2.mga8.x86_64
      1/4: removing virtualbox-kernel-5.15.62-desktop-1.mga8-6.1.38-1.2.mga8.x86_64
                                 ###########################################################################

-------- Uninstall Beginning --------
Module:  virtualbox
Version: 6.1.36-1.mga8
Kernel:  5.15.58-desktop-2.mga8 (x86_64)
-------------------------------------

Status: Before uninstall, this module version was ACTIVE on this kernel.

vboxdrv.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.58-desktop-2.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.


vboxnetflt.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.58-desktop-2.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.


vboxnetadp.ko.xz:
 - Uninstallation
   - Deleting from: /lib/modules/5.15.58-desktop-2.mga8/dkms-binary/3rdparty/vbox/
 - Original module
   - No original module was found for this module on this kernel.
   - Use the dkms install command to reinstall any previous module version.

depmod.........

DKMS: uninstall Completed.
removing package virtualbox-kernel-5.15.58-desktop-2.mga8-6.1.36-1.4.mga8.x86_64
      2/4: removing virtualbox-kernel-5.15.58-desktop-2.mga8-6.1.36-1.4.mga8.x86_64
                                 ###########################################################################
removing package kernel-desktop-5.15.58-2.mga8-1-1.mga8.x86_64
      3/4: removing kernel-desktop-5.15.58-2.mga8-1-1.mga8.x86_64
                                 ###########################################################################
removing package kernel-desktop-5.15.62-1.mga8-1-1.mga8.x86_64
      4/4: removing kernel-desktop-5.15.62-1.mga8-1-1.mga8.x86_64
                                 ###########################################################################

(In reply to Morgan Leijström from comment #9)
> (In reply to Edward from comment #7)
> > (In reply to Morgan Leijström from comment #5)
> > I have been pressing the power button to power down and get the system back.
> 
> The interesting thing i asked about is if the system powers down after a
> short press, or you need to hold it until power cuts.
> 
> 
> > The RAM is fairly new (2020) and had been tested a while back, showing no
> > errors.
> 
> I also had new ram, but the memory tester also detect other problems in the
> data paths, such as connectors, and in my case it found a bad CPU core.
> 
> 
> Anyway, your comment 8, terminal working, is what we can proceed on.

There was one very early instance, where I did not hold in the power button long enough and with that, it rebooted without powering down.

Just had another Firefox freeze, however, was not running VirtualBox. It was on the host (Mageia 8).

Also had Chromium running at the same time, which did not freeze.

See if any of the steps in https://support.mozilla.org/en-US/kb/firefox-hangs-or-not-responding work.

(In reply to Dave Hodgins from comment #17)
> See if any of the steps in
> https://support.mozilla.org/en-US/kb/firefox-hangs-or-not-responding work.

The easiest appeared to be the deletion of the places.sqlite file, which recreates itself and restores the bookmarks upon the next launch.

It was fine for only a few days, then it froze again.

I have since deleted all of the non-Mageia VirtualBoxes *and so far*, the system has been stable, no mga8 freezes and Firefox has also behaved. All of the remaining VirtualBoxes are Cauldron, except for one on mga8.

But I don't understand how using a different Linux OS in a VirtualBox could cause this to occur.

Only thing I can think of is a hardware problem such as a bad spot on a
disk drive that one or more of those vb guests were triggering. Either that
a bug in vb triggered by one or more of those guests that Mageia doesn't
trigger.

At this point in time, I think an HDD problem can be ruled out. I just ran GSmartControl and performed a Short self-test on the drive. It reported no errors. The last Extended self-test was run only 520 lifetime hours ago and that also reported no errors. I also installed a Mageia 8 VBox just a few days ago, I would guess that it would have saved that to the HDD using space where the deleted VBoxes were previously saved.

I would like to further update this report, as an unrelated issue (without a bug report) has apparently been resolved by using a different desktop environment (DE) on the host than what I had been using, which was Xfce.

I will be using this other DE and running each VirtualBox, to see if mga8 still freezes up while performing the same actions that caused the freezes.

mg8 froze using the other desktop (LXDE) as the host, so I now believe the problem is an incompatibility between the kernel and VirtualBox kernel modules.

I agree that a disc fault is extremely unlikely, especially as you have run the SMART self-tests. I recently saw a bad disc where, after running the long self-test, Gsmartcontrol on starting immediately warned of imminent failure, panick!

CC'ing Thomas because of your latest suspicion above. For his view.

CC: (none) => lewyssmith, tmb

Edward
Is this still an issue?
Having perused the entire bug, I was confused about exactly what is freezing:
- without any virtual machines, a normal system? M8?
- with virtual machines:
-- something on the host system? or the entire host system? M8?
-- something on the guest system? or the entire guest VM?  M8 or M9?

Technically, no. 

I no longer use VirtualBox due to various issues. Plus, I'm using KDE exclusively now.

Thanks for that.

Status: NEW => RESOLVED
Resolution: (none) => OLD