[#ABICLOUD-688] VSM does not unsunscribe more than 1 VM in Hyper-V after undeploying a VApp - ABIQUO

Abiquo

VSM does not unsunscribe more than 1 VM in Hyper-V after undeploying a VApp

Created: 30/Aug/10 11:00 AM Updated: 17/Sep/10 11:35 AM Resolved: 17/Sep/10 11:19 AM

Component/s:

Virtualization

Affects Version/s:

Abiquo-1.6.0

Fix Version/s:

Abiquo-1.6.5

Time Tracking:

Original Estimate:

4 hours

Remaining Estimate:

4 hours

Time Spent:

Not Specified

Affect Build:

3361

Description

« Hide

Steps to Reproduce
-enter as sysadmin
-go to virtual datacenter
-select a virtual datacenter and virtual appliance with hyper-v hypervisor
-add several image to the Virtual appliance and deploy -> vsm subscribe to changes for every VM
-undeploy

Expected Results
When undeploying, vsm should unsuscribe to every VM

Real Results
VSM only unsuscribe to the first Virtual Machine in virtual appliance. For the others ones, is displayed in log each 10 seconds:
10:58:00.108 WARN c.a.a.vsm.plugin.hyperv.HyperVPoller - Could not get state of Virtual System: ABQ_db224c7e-996f-4ba8-80e6-bcbaa1075141
10:58:05.775 WARN c.a.a.vsm.plugin.hyperv.HyperVPoller - Could not get state of Virtual System: ABQ_db224c7e-996f-4ba8-80e6-bcbaa1075141
10:58:11.451 WARN c.a.a.vsm.plugin.hyperv.HyperVPoller - Could not get state of Virtual System: ABQ_db224c7e-996f-4ba8-80e6-bcbaa1075141

Sort Order:

[ Permalink | « Hide ]

Ignasi Barrera added a comment - 02/Sep/10 02:31 PM

Fixed in revision 3378.

[ Permalink | « Hide ]

Valerie Kalusinski added a comment - 14/Sep/10 05:58 PM

this has been fixed for several VM running in same Virtual Appliance.
But we have the same problem with removed captured VM.

Steps to Reproduce
-enter as sysadmin
-go to infrastructure
-select a datacenter, physical machine hyper-v and capture a VM
-go to hyper-v and remove the captured VM
-go to the Virtual appliance with the removed captured VM
-undeploy

-> vsm does not suscribe to the captured removed VM.

Reopened on build 3412.

[ Permalink | « Hide ]

Ignasi Barrera added a comment - 17/Sep/10 11:19 AM

This is a normal behavior:

The HyperV monitor is implemented as a poller that asks for the state of the VM every X seconds.
When a VM is removed directly from the Hypervisor, the Monitor has not enough information to decide if it must be unregistered: It doesn't know if the VM has disappeared, if there is a punctual networking problem that makes it impossible to monitor the VM, etc, so a trace is written in the log to inform about the situation.

[ Permalink | « Hide ]

Valerie Kalusinski added a comment - 17/Sep/10 11:35 AM

Verified.