Ubuntu Guest Loses Network Connection

[jasblade]

Good day, I'm looking for some guidance or solutions for this issue I am seeming to have.
For my issue I'm using:
Windows 7 (32bit) host
Ubuntu 32bit (either 10.04 or 10.10) guest OS'
VirtualBox 4.02 OR 3.2
--Guest is allocated 1024mb mem
--Guest is allocated 64mb of video mem
--Tested with and without cpu extended features PAE/NX
--Tested with and without accleration for VT-x/AMD-V and Nested Paging
--Tested with and without 3D Acceleration


Background: These were guest OS' I ran before we were using Win XP sp3 and the guest OS' operated flawlessly with no issues (except using advanced 3d graphics like compiz, but I really do not need to use anything fancy to get my work done)

Moved to windows 7 and vbox 4.0.2 was installed so I copied my .vdi over and it worked beautifully for 3 days. Last Saturday I started my day with the 10.04 OS running for about 10-15 minutes than I lost my SSH tunnel to another system.
--Running ifconfig I find 0 errors, drops, etc
--Running ethtool I show link connected, full duplex 1gb/s
When I check my routing with route -v I show no routes, I'm unable to reach to vbox's dhcp (which is how its running with NAT from the host, as per default configs)

I'm completely unable to reach dhcp or reach out of vm if I change everything to static and I still have problems reaching out.

I have not yet added a network adapter and tried to create a bridge as I'm not really sure on the procedure in doing that.

Here's part of the how the problem got worse for me:
If I reboot the vm completely and check vbox logs I really am not seeing any indication of what is going on that might be wrong, but if I restart the virtualbox application and reboot the guest I will once again get about 3-15 mins of use before I lose my network connections. I'm also seeing my guest OS keep identifying new adapters (eth1..eth4 now).

If I reboot the guest enough, it finally gets to a state of only kernel panicking on boot and will not boot anymore.

If I take the same .vdi to my Win 7 (64bit) Host at home, the guest boots and runs flawlessly even with seemless mode and some other more advanced GA features.

Any suggestions?
Thanks --Jason

[jasblade]

Just giving an update. A coworker here was having panics with CentOS (latest greatest, but not sure of exact build) and Ubuntu 10.10 32bit. He said he finally got a new install to work using Vbox 3.2.12. He's using identical Win7 32bit host system.

I uninstalled vbox, rebooted installed 3.2.12 and tried my 10.04 that I've used for quite awhile and it launched and worked fine for about two hours. I am also running the matching GA for 3.2.12. As per my typical issue, I lost my network connection to remote systems and when checking my NIC with ifconfig, it shows no issues. Ethtool shows full duplex and link, however ethtool -t fails it's test and the Network-Manager officially tells me my link is lost. If I run a shutdown on the guest, the vbox manager seems to be unable to cleanly close the guest out and is having issues with leaving open file handles to the vbox.log file.

I end up with one or more VBoxSVC.exe processes in my host that simply can not be ended. I can force the pid off the tasklist with a taskkill /F from the cli but it seems something in os/mem/cpu is holding onto a file handle. The typical error references a fopen failing. I can't seem to reproduce this exact error at the moment, however this time I am looking at:

Code: Select all

Failed to open a session for the virtual machine vubuntu
Failed to assign the machine to the session (Unknown Status 0x80004005)

Result Code: 
VBOX_E_VM_ERROR (0x80BB0003)
Component: 
Machine
Interface: 
IMachine {6d9212cb-a5c0-48b7-bbc1-3fa2ba2ee6d2}
Callee: 
IVirtualBox {3f36e024-7fed-4f20-a02c-9158a82b44e6}
Callee RC: 
E_FAIL (0x80004005)

This appears after I have closed the vbox manager and restarted it, the state of the VM does show "Powered Off" not italicized as it did prior to closing it.
The end of my current VBox.log file shows:

Code: Select all

02:42:43.078 NAT: zone(nm:mbuf, used:201)
02:42:43.079 NAT: zone(nm:mbuf_cluster, used:177)
02:42:43.079 NAT: zone(nm:mbuf_packet, used:0)
02:42:43.079 NAT: zone(nm:mbuf_jumbo_pagesize, used:0)
02:42:43.079 NAT: zone(nm:mbuf_jumbo_9k, used:0)
02:42:43.079 NAT: zone(nm:mbuf_jumbo_16k, used:23)
02:43:43.099 Changing the VM state from 'DESTROYING' to 'TERMINATED'.

This is contrary to the previous log showing (which is repeated over a hundred times)

Code: Select all

01:07:24.031 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={f6916ec5-a881-4237-898f-7de58cf88672} aComponent={Keyboard} aText={The console is not powered up}, preserve=false
01:07:24.095 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={f6916ec5-a881-4237-898f-7de58cf88672} aComponent={Keyboard} aText={The console is not powered up}, preserve=false
01:07:24.167 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={f6916ec5-a881-4237-898f-7de58cf88672} aComponent={Keyboard} aText={The console is not powered up}, preserve=false
01:07:24.287 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={f6916ec5-a881-4237-898f-7de58cf88672} aComponent={Keyboard} aText={The console is not powered up}, preserve=false
01:07:32.418 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={09eed313-cd56-4d06-bd56-fac0f716b5dd} aComponent={Display} aText={The console is not powered up}, preserve=false
01:07:32.453 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={515e8e8d-f932-4d8e-9f32-79a52aead882} aComponent={Console} aText={The virtual machine is being powered down}, preserve=false
01:07:32.503 ERROR [COM]: aRC=E_ACCESSDENIED (0x80070005) aIID={515e8e8d-f932-4d8e-9f32-79a52aead882} aComponent={Console} aText={The virtual machine is being powered down}, preserve=false
01:07:32.518 ERROR [COM]: aRC=VBOX_E_INVALID_VM_STATE (0x80bb0002) aIID={515e8e8d-f932-4d8e-9f32-79a52aead882} aComponent={Console} aText={The virtual machine is already running or busy (machine state: Stopping)}, preserve=false

[/size]

Hoping someone can point me in a better direction for keeping my network connection alive in the guest, as if I can't, I'm reduced to a couple dozen putty windows, no screen/tmux, no clusterssh, and no tunnel past the proxy for gtalk, icanhazcheezburger, half a dozen technical forums, and all my lovely profile settings all of us love to nerd-rage about when they get taken away.

[jasblade]

anyone ?

[asf]

I am having the same problems. Things have been working fine with XP SP3 as a host system. The same VM (Ubuntu 10.10) running on Win7 Pro repeatedly loses connectivity after some time.

Unfortunately we haven't been able to find a solution for this issue either. Any pointers or help would be appreciated. Thanks.

[jasblade]

I have some comfort knowing I'm not alone. This same .vdi of 10.04 and the one for 10.10 work without issues on my Win7 64bit host system.
The chipset in this 32bit Win7 machine shows:

Intel(R) 82578DM Gigabit Network Connection

The rest of the motherboard is the Q57 Express / 3400 series Intel stuff.
cpu is a 3.33ghz i5.

Is this the same thing you have?

If this was a system that didn't have to stay matching an image I would try generic NIC drivers

[rootkit]

same thing here. I use Ubuntu 10 as host and two windows XP as guests. The connections on guests lose connection like one time each day, I have to shut down the guests and give them new MAC to get network work back again. Please help...

[rootkit]

any help on this issue is appreciated !!

[vijaymverma]

Since today morning my Ubuntu guest connection also is non-responsive. No issues with NAT, no IP issues, there is just no network traffic. It seems strange. Hopefully there is a fix or I will try to roll back to 4.0.2 and see how it goes.

[jasblade]

I still continue to have the problem since upgrading to VB v4.0.4
Additionally I have a problem not being able to boot my VM unless I allocate 2cpus and all the acceleration options to the VM. With a single cpu is simply kernel panics.

Booting with 2 cpus and acceleration it boots very quickly and operates smoothly, but I can't seem to get more than an hour of uptime and the entire VM appears to lock up. In some cases I can still move the term windows around on my desktop for a min before they two stop responding. There usually is no way I can properly close down the OS when this happens.

If I reboot the VM after task manager kills the vbox service exe process the VM will start, but usually in less then 2mins loses it's network connectivity. I can usually close down the VM properly when this happens.

In order to get things working for my hour of work I can get done, I usually need to reboot my host windows os first. But, same pattern keeps happinging, and I'm not able to find any logs or events that can give a clue as to why or how to create a work around.

At this point I am even tired of kicking puppies to deal with this :'(

[faflu]

The same issue on win XP SP3(host) + Xubuntu 10.10 (guest). For almost 2 years I haven't had such issue with VBox, and then all of the sudden 2 days ago it started...
I need to shut down the guest system and boot it again to get network back.
There were no changes in the host or guest systems. Just suddenly it stopped working... I also experience crashes of virtualbox engine - "the program has generated an exception and will be closed"....
Vbox has been reinstalled, no progress.