Teleporting broken (VBox 4.1.x)

[lostogre]

I am running VB 4.1.2 on two Ubuntu Lucid Linux 10.04.3 hosts with iSCSI shared storage and 26GB RAM per host. The guests are a mix of WindowsXP, Windows2003, and mostly Ubuntu 10.04.3.

When I try to teleport a VM on one host to the other host, something that worked correctly on 4.0.12, I get a guru meditation error.

[Frank Mehnert]

Yes, teleporting with 4.1 seems to be broken. Current workaround is to stay at VBox 4.0.

[Nik]

Additional info: The sending end of the setup gets this error:

D:\Documents and Settings\autotest>"C:\Program Files\Oracle\VirtualBox\vboxmanage" controlvm  ta-mini teleport --hostnam
e automation2 --port 6432
0%...
Progress state: E_FAIL
VBoxManage.exe: error: Teleportation failed. Error message: VMR3Teleport -> VERR_SSM_LIVE_GURU_MEDITATION

The receiving end of the setup gets:
C:\Program Files\Oracle\VirtualBox\vboxmanage.exe" modifyvm ta-mini --teleporter off

VBoxManage.exe: error: The machine 'ta-mini' is already locked for a session (or being unlocked)
VBoxManage.exe: error: Details: code VBOX_E_INVALID_OBJECT_STATE (0x80bb0007), component Machine, interface IMachine, callee IUnknown
Context: "LockMachine(a->session, LockType_Write)" at line 367 of file VBoxManageModifyVM.cpp

[Nik]

Correction to receiving end command log:

"C:\Program Files\Oracle\VirtualBox\vboxmanage.exe" modifyvm ta-mini --teleporter on
--teleporterport 6432

"C:\Program Files\Oracle\VirtualBox\vboxmanage.exe" startvm ta-mini
Waiting for VM "ta-mini" to power on...

D:\Documents and Settings\autotest>"C:\Program Files\Oracle\VirtualBox\vboxmanage.exe" modifyvm ta-mini --teleporter on
--teleporterport 6432

"C:\Program Files\Oracle\VirtualBox\vboxmanage.exe" startvm ta-mini
Waiting for VM "ta-mini" to power on...
VBoxManage.exe: error: Failed to load unit 'pgm' (VERR_SSM_CANCELLED)
VBoxManage.exe: error: Details: code E_FAIL (0x80004005), component Console, interface IConsole, callee

[Twisted Lucidity]

I am seeing this as well with 4.1.8 on Windows 7 64bit and Ubuntu 10.10 64bit hosts.

It happens when trying to teleport between hosts, and when trying to test by teleporting to localhost.

The guests I have tried have all been Linux.

Dialog message:

Failed to open a session for the virtual machine dev-lubu-ji.
Failed to load unit 'pgm' (VERR_SSM_CANCELLED). 

Result Code: 
NS_ERROR_FAILURE (0x80004005)
Component: 
Console
Interface: 
IConsole {1968b7d3-e3bf-4ceb-99e0-cb7c913317bb}

Console:

Progress state: NS_ERROR_FAILURE
VBoxManage: error: Teleportation failed. Error message: VMR3Teleport -> VERR_SSM_LIVE_GURU_MEDITATION

Sending log:

00:01:43.731 SSM: Execute save failed with rc=VERR_SSM_CANCELLED for data unit 'pgm'/#1.
00:01:43.731 SSM: Failed to save the VM state.
00:01:43.742 Changing the VM state from 'GURU_MEDITATION_LS' to 'GURU_MEDITATION'.
00:01:43.743 ERROR [COM]: aRC=NS_ERROR_FAILURE (0x80004005) aIID={1968b7d3-e3bf-4ceb-99e0-cb7c913317bb} aComponent={Console} aText={VMR3Teleport -> VERR_SSM_LIVE_GURU_MEDITATION}, preserve=false

Receiving log:

00:01:52.832 SSM: LoadExec failed for 'pgm' instance #1 (version 14, pass 0x1): VERR_SSM_CANCELLED
00:01:52.832 VMSetError: /home/vbox/vbox-4.1.8/src/VBox/VMM/VMMR3/SSM.cpp(8155) int ssmR3LoadExecV2(VM*, SSMHANDLE*); rc=VERR_SSM_CANCELLED
00:01:52.832 VMSetError: Failed to load unit 'pgm'
00:01:52.832 Changing the VM state from 'LOADING' to 'LOAD_FAILURE'.
00:01:52.832 Teleporter: VMR3LoadFromStream -> VERR_SSM_CANCELLED
00:01:52.832 ERROR [COM]: aRC=NS_ERROR_FAILURE (0x80004005) aIID={1968b7d3-e3bf-4ceb-99e0-cb7c913317bb} aComponent={Console} aText={Failed to load unit 'pgm' (VERR_SSM_CANCELLED)}, preserve=false

[Twisted Lucidity]

The Receiving VM's log file

[Twisted Lucidity]

The Sending VM's log file

[WK]

We are also waiting to get this fixed. 4.0.x is getting long in the tooth and we would like to move into 4.1.x but teleport is too useful to lose.

[schmidtc]

The Bug is still present in VirtualBox 4.2.0. I have attached the Log files of a failed migration on a Solaris 10 AMD64 Node.

[Frank Mehnert]

Correct, the bug was not yet fixed. But the priority for fixing this bug was now a bit increased. Though still no ETA for a fix, sorry.