Opened 8 years ago
Closed 7 years ago
#15869 closed defect (duplicate)
VBoxNetNat.exe: The instruction at 0x referenced memory at 0x. The memory could nom be written.
| Reported by: | the_real_spiffytech | Owned by: | |
|---|---|---|---|
| Component: | other | Version: | VirtualBox 5.1.4 |
| Keywords: | Cc: | ||
| Guest type: | Linux | Host type: | Windows |
Description
I often (unpredictably) get this error that kills my VM's networking until I restart or save/restore:
The instruction at 0x referenced memory at 0x. The memory could nom be written. VBoxNetNat.exe
I can't find a deliberate trigger, though it may possibly be related to network throughput. Or talking to a Kubernetes cluster.
Attachments (7)
Change History (18)
by , 8 years ago
| Attachment: | vbox_log.log added |
|---|
comment:1 by , 8 years ago
Very likely that these crashes are fixed in the most recent 5.1.x Windows test build which you can find here. Could you test? Thank you!
comment:2 by , 8 years ago
I've installed the test release and will report back if I encounter the error. Since the error has not been deliberately reproducible, I won't be able to provide an immediate response.
comment:3 by , 8 years ago
Negative: I got the same error again. Plus, on a separate occasion, a new problem with the same symptoms (VM networking dies unless restarted), just without the error message.
New logs + minidumps attached.
comment:4 by , 8 years ago
I get similar message, on all the latest win test builds (win7)when I run ubuntu 16 04 vm, updated using apt-get. Happens just as vm is about to start the gui. The last test build on which this vm runs ok was 110354 - ie just few days ago. Also ok on 5.1.4 release.
Text in the window is: instruction 0xe82fe4b6 cannot write memory 0x0000000018
and then the vm crashes & cannot be started
john
comment:5 by , 8 years ago
Mine is a Windows 7 host, Debian Jessie up-to-date. I can successfully operate the GUI indefinietly, and had this problem since VirtualBox v5.0.something.
When the message pops up, if I dismiss it by hitting Esc I can safely shut down / suspend the VM, and it restarts fine.
comment:6 by , 8 years ago
the_real_spiffytech, thanks for the crash dump. Am I right that this VBoxNetNAT crash happens occasionally and it happens during normal networking? Is IPv6 traffic involved? According to the crash dump it might crash in IPv6 code.
comment:7 by , 8 years ago
I'm not knowingly connecting to IPv6 hosts. But these days, that doesn't mean much.
I was out most of the day on an IPv4-only network, and I think I had this issue a couple times early in the day before you suggested the IPv6 possibility, but all the instances are bleeding together.
I came home to my IPv6 network and see my VM isn't picking up an IPv6 address, even though my host passes http://test-ipv6.com.
Can you suggest tests I could run, or a way to detect potentially-triggering IPv6 activity?
comment:8 by , 8 years ago
Actually it's very hard to tell if the crash is really related to IPv6 or not. There is a crashing thread which seems to be related to some IPv6 activity but OTOH the stack of this thread seems to be corrupted.
Could you try to find out how to best reproduce the crash? Are you browsing some certain web page, any special network protocol (VPN, ssh, etc)? How long do you need to reproduce the crash?
comment:9 by , 8 years ago
I have been keeping my eye out for a couple months, but haven't seen anything consistent. The only thing I've noticed is an increased likelihood of crashes when I'm doing work with the Kubernetes HTTP API. But even that doesn't trigger it consistently, and I haven't found a way to narrow it down further.
Sometimes it's 6 crashes/day, sometimes it's 6 days/crash. Very inconsistent. Makes me think it's triggered by something specific I'm doing some days and not others, but no luck identifying the culprit yet.
vbox.log