Faulty teleportation configuration leaves half-open VBox

[bauer40]

I know that what I've done is not allowed, but it should be handled correctly...

If you configure teleportation on a port which is in use, and start up VBoxHeadless, the VM is being shut down but the state of the VM is only "stopping", not "powered off".

To reproduce, create two vms, and issue:

VBoxManage modifyvm testvm1 --teleporter on --teleporterport 6000
VBoxManage modifyvm testvm2 --teleporter on --teleporterport 6000 <- that's the problem

nohup VBoxHeadless -s testvm1 <- testvm waits for teleportation
VBoxheadless -s testvm2 <- This one is terminated

When you look on the VBox GUI, the state of testvm2 is "stopping", which is wrong. Furthermore, you can no longer start/stop testvm2.

[bauer40]

Update: The problem seems to be that only the GUI is not updated.

Sorry for assigning the ticket to the wrong group.

[Ramshankar Venkataraman]

Is this still an issue with the latest VirtualBox?

[bauer40]

Replying to ramshankar:

Is this still an issue with the latest VirtualBox?

Yes and no. The behaviour is better, but still not the best (in my opinion). However, the system is kept usable.

When reproducing the problem with the above mentioned method, the start of the second machine fails (because port 6000 is already in use) and the status of the VM is set to "Powered Off" (which is quite safe). But VBoxHeadless does not return to the shell with an error code, as I would expect. VBoxHeadless keeps running until Control-C'ed, leaving the VM "aborted".

Here is the output:

bash-3.00# VBoxHeadless -s testvm2
Oracle VM VirtualBox Headless Interface 3.2.8
(C) 2008-2010 Oracle Corporation
All rights reserved.

Listening on port 7002.
VRDP server is inactive.
Error: failed to start machine. Error message: RTTcpServerCreateEx failed with status VERR_NET_ADDRESS_IN_USE

''VBoxHeadless does not return to the shell, making it impossible to handle that situation in a script. So, I have to cancel''
^C
bash-3.00#

[bauer40]

VBox.log of 3.2.8 after starting up a VM which cant be started because the teleporatation port is in use. Therefor, I have to Control-C the process.

[Klaus Espenlaub]

The problem you reported is completely unrelated to the original problem. 3.2.8 has a regression with starting VMs with broken config, the hang you observed. This can happen for many reasons, and your duplicate port with teleportation is just one of them.

[bauer40]

OK, in that case the original issue seems to be resolved.

What do you think about the fact that VBoxHeadless does'nt come back to the shell in the discribed situation? Should I document a bug?

[Ramshankar Venkataraman]

No, don't think it's necessary as the issue has just been fixed and will be available in the next release.

[Ramshankar Venkataraman]

Closing as the original issue is resolved.