VirtualBox

Opened 10 years ago

Closed 10 years ago

#12861 closed defect (worksforme)

general protection fault: 0000 [#1] SMP

Reported by: Ken Preslan Owned by:
Component: other Version: VirtualBox 4.3.8
Keywords: Cc:
Guest type: Linux Host type: Linux

Description

I'm running VBox on a 64-bit Ubuntu 12.04 host. It's running three VMs of the same version of Ubuntu. There are two ethernet interfaces. One VM is bridged onto both. The other two are bridged onto just eth0.

I've been running VBox in the configuration for years with no problems. After I upgraded to 4.3.8, I've had the machine spontaneously reboot twice. This evening, I had it panic with the attached message in the logs. The stack back-trace was incomplete.

I'll try to install kdump and hopefully get a core and better information next time.

I'm not sure this is VBox's fault, but right now it seems like the most likely culprit.

Is there any other information I should provide?

Attachments (4)

VBox.log.1 (97.9 KB ) - added by Ken Preslan 10 years ago.
From the VM that's bridged to both interfaces. The logs for all three VMs have the part about "VirtualBoxClient: detected unresponsive VBoxSVC (rc=NS_ERROR_ABORT)".
gpf.txt (3.5 KB ) - added by Ken Preslan 10 years ago.
The partial GPF message that showed up in /var/log/kern.log.
malloc_fail.txt (6.5 KB ) - added by Ken Preslan 10 years ago.
I've always gotten messages like this occasionally. I'm not sure if it's related.
crash_dmesg.txt (9.5 KB ) - added by Ken Preslan 10 years ago.
The crash text taken from a kdump when it happened again. It doesn't seem to be a smoking gun either way, but the crash is in VBoxXPCOMIPCD again.

Download all attachments as: .zip

Change History (14)

by Ken Preslan, 10 years ago

Attachment: VBox.log.1 added

From the VM that's bridged to both interfaces. The logs for all three VMs have the part about "VirtualBoxClient: detected unresponsive VBoxSVC (rc=NS_ERROR_ABORT)".

by Ken Preslan, 10 years ago

Attachment: gpf.txt added

The partial GPF message that showed up in /var/log/kern.log.

by Ken Preslan, 10 years ago

Attachment: malloc_fail.txt added

I've always gotten messages like this occasionally. I'm not sure if it's related.

by Ken Preslan, 10 years ago

Attachment: crash_dmesg.txt added

The crash text taken from a kdump when it happened again. It doesn't seem to be a smoking gun either way, but the crash is in VBoxXPCOMIPCD again.

comment:1 by Ken Preslan, 10 years ago

I upgraded to 4.3.10 (92957) last night and installed kdump. This afternoon I got another crash. I attached the error text I got by running dmesg inside the crash program.

I can upload the dump if someone wants to see it.

I downgraded to 4.3.6 again. I'll run that way for a while and add a comment if I see a crash again or not.

Thanks!

comment:2 by Frank Mehnert, 10 years ago

I would be very interested in the kdump. Is it less than 10MB (compressed) so you could email me the dump? frank _dot_ mehnert _at_ oracle _dot_ com ... Otherwise, could you make the image available to me? Unfortunately the anonymous FTP server is not available anymore :-/

comment:3 by Ken Preslan, 10 years ago

I sent Frank an email with a link to the kdump.

comment:4 by Ken Preslan, 10 years ago

I had a crash with 4.3.6. Again in VBoxXPCOMIPCD.

Given that I ran with 4.3.6 since it came out with the same config, I'm really starting to doubt this is VirtualBox's fault. I think my hardware is subtly screwed somehow.

Sorry for the noise.

comment:5 by Ken Preslan, 10 years ago

Memtest86 passes. The diagnostics in the machine's bios pass.

I downgraded the host to the 3.2.0-59 kernel and the machine hasn't crashed since. It's still a bit early to claim that's the culprit, but I'm hoping.

comment:6 by Ken Preslan, 10 years ago

D'oh. Another crash with 3.2.0-59.

I'm not sure what to try next.

comment:7 by Frank Mehnert, 10 years ago

Analyzed the kdump. Unfortunately I see a wrong access which could be due to memory corruption. On the other hand faults seem to happen always during kmalloc.

Any word about how easy to reproduce the problem? You said that this started after updating VirtualBox to 4.3.8. Which was the version you used before?

comment:8 by Ken Preslan, 10 years ago

Thanks for taking the time to look at my problem.

I don't think it would be easy to reproduce. My machine crashes every couple of days, but there's no specific thing I do to trigger it. In fact, it's never crashed while I was interacting with it. I can wail on it pretty hard and it runs fine. Then I leave it and do something else and it crashes while I'm away.

The machine is a little server in my basement and runs an apache VM, a postfix/dovecot VM, and a VM that does DHCP/NAT for my house. So, it's never completely idle, but it only has a lot of load when I'm doing something to it.

I've turned off the DHCP server on the last VM and have my house's traffic going through a different box. (Just so my house's internet connection doesn't get disrupted every time the machine crashes.) The machine does a lot less, and it has only crashed once since. So, system load is somewhat of an issue, but I still can't get it to crash at will.

I set up my older 32-bit, non-vmx box in a similar configuration and ran that for a while. I didn't see a crash, but I'm not sure if that's just because the hardware was so much slower. Or maybe it points to a hardware issue. I'm not sure.

I updated from 4.3.6 to 4.3.8 a few days before the first crash. Before that I was running 4.3.4. I generally skip the .0 and .2 of any new release, but other than that I've been running the lastest VirtualBox on this machine since I bought it 3.5 years ago.

I've reproduced the crash with 4.3.6, 4.3.8, 4.3.10, and 4.2.24.

Thanks!

comment:9 by Ken Preslan, 10 years ago

I just had the panic without any VirtualBox software on the machine at all.

Invalid bug.

Sorry.

comment:10 by Frank Mehnert, 10 years ago

Resolution: worksforme
Status: newclosed

Thanks Ken for letting us know!

Note: See TracTickets for help on using tickets.

© 2023 Oracle
ContactPrivacy policyTerms of Use