VBox 4.3.26 NAT Network unstable

[tkvb]

VBox 4.3.26 running on Windows 8.1 x64 host 2 Ubuntu 14.04.2 VMs configured to use a NAT network

After a while (maybe under heavy network load between the 2 VMs?) the NAT network stops passing traffic to the outside world. The 2 VMs can still talk to each other but not with the outside world until they are both rebooted.

This started with the upgrade to 4.3.26, the same setup worked OK before.

[Valery Ushakov]

Please, can you attach the log file for the NAT Network (.VirtualBox/$netname.log).

[tkvb]

I attached the log file with the VMs in the broken state, but there isn't much in the log file to go by. Is there a way to bump up the log level to get more info?

[tkvb]

I just got the following VBoxNetNAT.exe crash, but there isn't any more info in the log file than before.

[Valery Ushakov]

This started with the upgrade to 4.3.26, the same setup worked OK before.

What was the previous release that you used. I don't think NAT Network had any non-trivial changes on 4.3 branch since last August.

Is there a way to bump up the log level to get more info?

Not without using a debug build.

I just got the following VBoxNetNAT.exe crash

Is this reproducible? If it is, please, could you provide a minidump?

[tkvb]

I used 4.3.22 and NAT worked fine (though that version had other issues non-NAT related). Then I upgraded to 4.3.24 and NAT also worked but DNS was broken (see ticket #13915). Now this problem in .26

I'm sorry but I disagree with the statement that only trivial changes are made in minor releases. It's more like 1 step forward, 2 backwards. I haven't been able to find a stable version of VBox in a long time, where the features I use just work.

(As a side note, another non-trivial change that was made in a patch release, I believe from .20 to .22, was to completely ignore %HOME% on Windows -- you simply don't make such changes in a patch release. See ​http://semver.org/)

As for the crash, I haven't found a pattern to reproduce it. The error window is very rare, but the NAT network stops working often and there's no new info added to the log file.

If you can provide a debug build or a build with a way to turn on/off verbose NAT logging, I'm more than willing to run it and help you find the issue.

[Reiner Brodbeck]

I also experience NATnetwork stop working in VBox 4.3.26

I use a Win7 SP1 host. I have a total of 7 VMs configured into two different NATnetworks. It seems that, if after stopping a VM there is no VM left in one of the NATnetworks, both NATnetworks stop working, i.e. a ping to the router (usually address x.x.x.1) does no longer work.

After stopping all VMs and restarting them all works fine again.

This configuration worked without any problem in VBox 4.3.22. I did not use 4.3.24 because of the DNS problem.

[Valery Ushakov]

Replying to ReinerB:

I also experience NATnetwork stop working in VBox 4.3.26

It seems that, if after stopping a VM there is no VM left in one of the NATnetworks, both NATnetworks stop working, i.e. a ping to the router (usually address x.x.x.1) does no longer work.

Please, can you file this as a separate bug report? I can reproduce this issue, but this bug report is about VBoxNetNAT.exe crash. Thanks.

[Valery Ushakov]

Replying to tkvb:

If you can provide a debug build or a build with a way to turn on/off verbose NAT logging, I'm more than willing to run it and help you find the issue.

Any luck with getting a minidump?

Having at least the first idea about where the crash might be would be helpful. Carpet-bombing with full debug logs is a bit of an overkill.

[tkvb]

I have not seen the crash lately, as mentioned before it's rare, but NAT networking still stops frequently. When it does I looked at the running processes and there are still 3 VBoxNetNAT processes running, so they have not crashed.

[Valery Ushakov]

Replying to tkvb:

I have not seen the crash lately, as mentioned before it's rare, but NAT networking still stops frequently. When it does I looked at the running processes and there are still 3 VBoxNetNAT processes running, so they have not crashed.

May be you can use task manager or ​Process Explorer to forcibly obtain a dump of the NAT network process that appears stuck like that? Of the 3 processes you want the grandchild, the one with the largest memory.

[tkvb]

I created a full dump of the process in the state where it's not passing any traffic to the outside world, but it's 84MB big, can you please contact me for a way to get it to you? Not sure I want to post it in the public, I don't know what private info there may be in there.

[Valery Ushakov]

Thanks. You can reach me at valery dot ushakov at oracle dot com

[tkvb]

Sent you an email, thanks.

[GoodGodd]

Hi there, im getting problem with my natnetwork its give me an error every 2 or last time 4 days, than its crash and people can't login to my server witch im running in VM. People who did'd logout they can still play but new players can't login to server.

When i will get this error again i will post here. If you can have a look in this please. Many thanks in advance.

This is logs from that daya.

VBoxSVC.log.1

VirtualBox COM Server 5.0.0 r101573 win.amd64 (Jul 9 2015 11:08:16) release log 00:00:00.015600 main Log opened 2015-08-05T13:58:57.783564800Z 00:00:00.015600 main Build Type: release 00:00:00.015600 main OS Product: Windows 7 00:00:00.015600 main OS Release: 6.1.7601 00:00:00.015600 main OS Service Pack: 1 00:00:00.046800 main DMI Product Name: PowerEdge R210 II 00:00:00.046800 main DMI Product Version: 00:00:00.046800 main Host RAM: 16353MB total, 14907MB available 00:00:00.046800 main Executable: C:\Program Files\Oracle\VirtualBox\VBoxSVC.exe 00:00:00.046800 main Process ID: 3496 00:00:00.046800 main Package type: WINDOWS_64BITS_GENERIC 00:00:00.046800 Home directory: 'C:\Users\ADMINMIR3/.VirtualBox' 00:00:00.046800 Loading settings file "C:\Users\ADMINMIR3/.VirtualBox\VirtualBox.xml" with version "1.14-windows" 00:00:00.358800 HostDnsMonitor: old information 00:00:00.358800 no server entries 00:00:00.358800 no search string entries 00:00:00.358800 no domain set 00:00:00.358800 HostDnsMonitor: new information 00:00:00.358800 server 1: 62.210.16.6 00:00:00.358800 server 2: 62.210.16.7 00:00:00.358800 no search string entries 00:00:00.358800 domain: online.net 00:00:00.358800 HostDnsMonitorProxy::notify 00:00:00.405600 VD: VDInit finished 00:00:00.405600 Loading settings file "C:\Users\ADMINMIR3\VirtualBox VMs\MIR 3\MIR 3.vbox" with version "1.14-windows" 00:00:05.538009 main ERROR [COM]: aRC=VBOX_E_OBJECT_IN_USE (0x80bb000c) aIID={4afe423b-43e0-e9d0-82e8-ceb307940dda} aComponent={MediumWrap} aText={Medium 'C:\Users\ADMINMIR3\VirtualBox VMs\MIR 3\MIR 3.vdi' cannot be closed because it is still attached to 1 virtual machines}, preserve=false aResultDetail=0 00:00:05.616009 Watcher ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={0169423f-46b4-cde9-91af-1e9d5b6cd945} aComponent={VirtualBoxWrap} aText={The object is not ready}, preserve=false aResultDetail=0

VBoxSVC.log.2

163:33:31.661974 Watcher Log continuation - Log started 2015-07-29T18:18:31.391260300Z 163:33:31.898004 Watcher Stopped NAT network 'NatNetwork' 163:33:40.258530 main ERROR [COM]: aRC=VBOX_E_OBJECT_IN_USE (0x80bb000c) aIID={4afe423b-43e0-e9d0-82e8-ceb307940dda} aComponent={MediumWrap} aText={Medium 'C:\Users\ADMINMIR3\VirtualBox VMs\MIR 3\MIR 3.vdi' cannot be closed because it is still attached to 1 virtual machines}, preserve=false aResultDetail=0 163:33:40.289730 Watcher ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={0169423f-46b4-cde9-91af-1e9d5b6cd945} aComponent={VirtualBoxWrap} aText={The object is not ready}, preserve=false aResultDetail=0

VBoxSVC.log.3

VirtualBox COM Server 5.0.0 r101573 win.amd64 (Jul 9 2015 11:08:16) release log 00:00:00.015600 main Log opened 2015-07-29T18:18:31.391260300Z 00:00:00.015600 main Build Type: release 00:00:00.015600 main OS Product: Windows 7 00:00:00.015600 main OS Release: 6.1.7601 00:00:00.015600 main OS Service Pack: 1 00:00:00.031200 main DMI Product Name: PowerEdge R210 II 00:00:00.031200 main DMI Product Version: 00:00:00.031200 main Host RAM: 16353MB total, 14970MB available 00:00:00.031200 main Executable: C:\Program Files\Oracle\VirtualBox\VBoxSVC.exe 00:00:00.031200 main Process ID: 3552 00:00:00.031200 main Package type: WINDOWS_64BITS_GENERIC 00:00:00.031200 Home directory: 'C:\Users\ADMINMIR3/.VirtualBox' 00:00:00.046800 Loading settings file "C:\Users\ADMINMIR3/.VirtualBox\VirtualBox.xml" with version "1.14-windows" 00:00:00.327600 HostDnsMonitor: old information 00:00:00.327600 no server entries 00:00:00.327600 no search string entries 00:00:00.327600 no domain set 00:00:00.327600 HostDnsMonitor: new information 00:00:00.327600 server 1: 62.210.16.6 00:00:00.327600 server 2: 62.210.16.7 00:00:00.327600 no search string entries 00:00:00.327600 domain: online.net 00:00:00.327600 HostDnsMonitorProxy::notify 00:00:00.358800 VD: VDInit finished 00:00:00.358800 Loading settings file "C:\Users\ADMINMIR3\VirtualBox VMs\MIR 3\MIR 3.vbox" with version "1.14-windows" 00:00:00.733201 ERROR [COM]: aRC=E_FAIL (0x80004005) aIID={4afe423b-43e0-e9d0-82e8-ceb307940dda} aComponent={MediumWrap} aText={Unknown exception 00:00:00.733201 F:\tinderbox\win-rel\out\win.amd64\release\obj\VBoxAPIWrap\MediumWrap.cpp[2245] (long cdecl MediumWrap::GetEncryptionSettings(unsigned short ,unsigned short ))}, preserve=false aResultDetail=0 00:00:06.416203 ERROR [COM]: aRC=E_FAIL (0x80004005) aIID={4afe423b-43e0-e9d0-82e8-ceb307940dda} aComponent={MediumWrap} aText={Unknown exception 00:00:06.416203 F:\tinderbox\win-rel\out\win.amd64\release\obj\VBoxAPIWrap\MediumWrap.cpp[2245] (long cdecl MediumWrap::GetEncryptionSettings(unsigned short ,unsigned short ))}, preserve=false aResultDetail=0 00:00:06.523717 ERROR [COM]: aRC=VBOX_E_OBJECT_NOT_FOUND (0x80bb0001) aIID={edba9d10-45d8-b440-1712-46ac0c9bc4c5} aComponent={ExtPackManagerWrap} aText={No extension pack by the name 'Oracle VM VirtualBox Extension Pack' was found}, preserve=false aResultDetail=0 00:00:06.877262 ERROR [COM]: aRC=VBOX_E_IPRT_ERROR (0x80bb0005) aIID={f30138d4-e5ea-4b3a-8858-a059de4c93fd} aComponent={SessionMachine} aText={Saved screenshot data is not available (VERR_NOT_SUPPORTED)}, preserve=false aResultDetail=0 00:00:07.246308 VM 'MIR 3' starts using NAT network 'NatNetwork' 00:00:07.364823 Started NAT network 'NatNetwork' 00:00:08.979028 HostDnsMonitorProxy::GetNameServers: 00:00:08.979028 name server 1: 62.210.16.6 00:00:08.979028 name server 2: 62.210.16.7 00:00:08.979028 HostDnsMonitorProxy::GetDomainName: online.net 00:00:09.214058 HostDnsMonitorProxy::GetNameServers: 00:00:09.214058 name server 1: 62.210.16.6 00:00:09.214058 name server 2: 62.210.16.7 00:00:44.282011 ERROR [COM]: aRC=E_FAIL (0x80004005) aIID={4afe423b-43e0-e9d0-82e8-ceb307940dda} aComponent={MediumWrap} aText={Unknown exception 00:00:44.282011 F:\tinderbox\win-rel\out\win.amd64\release\obj\VBoxAPIWrap\MediumWrap.cpp[2245] (long cdecl MediumWrap::GetEncryptionSettings(unsigned short ,unsigned short ))}, preserve=false aResultDetail=0 163:33:31.655473 Watcher VM 'MIR 3' stops using NAT network 'NatNetwork' 163:33:31.655473 Watcher Log rotated - Log started 2015-07-29T18:18:31.391260300Z

[Valery Ushakov]

Replying to tkvb:

I have not seen the crash lately, as mentioned before it's rare, but NAT networking still stops frequently. When it does I looked at the running processes and there are still 3 VBoxNetNAT processes running, so they have not crashed.

When your NAT Netowrk is stuck, please, can you do netstat -n -A inet -A inet6 in all your guests connected to that network?

I suspect you are exceeding the limit of simultaneous TCP connections. E.g. gvfsd-http apparently likes to keep around a lot of connections in CLOSE_WAIT (launchpad bugs ​571970 and ​760344).

[GoodGodd]

Vbox error

[GoodGodd]

Hi, hope anyone can help! We having this error few months now.

I'm getting same error, my Nat Network crash every 2 or 7 days. It's like this. We running game server on VBox 5, Host Windows 7 and guest Win Server 2003 and one of our software is for connection players to the server (gate soft - only for login to game using credentials). When we have 20+ people connected in one time NAT crashes every up to 7 days but when we having 40+ people connected in one time it crashes every 2 days. Please have a look on attached file error.png

So i wonder if we can fix it somehow. Many thanks in advance.

Regards Maciej

[Valery Ushakov]

Please, can you give a recent test build a try? Revision 103571+ for 5.0, 103575+ for 4.3.

[Frank Mehnert]

No response, closing.