You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cloudstack.apache.org by Sanjeev Neelarapu <sa...@citrix.com> on 2012/05/01 07:47:12 UTC

RE: Test cases for Open vSwitch Tunnel Manager

Adding  cloudstack-dev

FYI..

From: Salvatore Orlando
Sent: Monday, April 30, 2012 2:08 PM
To: Sanjeev Neelarapu; Kevin Kluge; Chiradeep Vittal
Cc: #Cloud - Bonita
Subject: RE: Test cases for Open vSwitch Tunnel Manager

Hi Sanjeev,

Thanks for preparing and sharing this document.

I have the following observations for you:
TEST #1,2,3,4 - The parameter is now called  "sdn.ovs.controller", still False by default
TEST #2,3 - Also, the range for Guest network ID (unfortunately still called VLAN range), should allow you to go beyond 4096
                Note: There's already a ticket for the labelling of the range (CS-14604)
TEST #5 - You can create the phy net with VLAN isolation, but if the flag is kept to true it will still try to create GRE tunnels
TEST #6 - [Probably trivial] Instance VM and the virtual router should be on two different hosts, so keep spinning up VMs until the first ends up on a host different from the one where the VR is deployed.
TEST #9 - Also verify that the vnet corresponding to the assigned GRE key is marked as 'taken' in the database
TEST #10 - I believe the only sequence of operations where this can work involves restarting the management server after switching off the sdn.ovs.controller flag. As far as I recall, this is a scenario we don't want to support, and therefore I would remove this test from the list. Please wait for Chiradeep's feedback for final confirmation.
TEST #14 - Tunnels across zones are not supported at the moment, and they are very likely to not be working at the moment. I'd remove this test as well, if Chiradeep agrees.
TEST #16 - It should verify that broadcasts received on ingress tunnel are not forwarded on egress tunnels, but are forwarded to VIF ports anyway.
TEST #17 - Same comment as for test #16
TEST #18,19,20 - OK, but probably not necessary as scenario is unsupported.
TEST #21 - How do we verify that the VM status is unpredictable?
TEST #26,27,28,29 - You'll need to ensure the VM being stopped is the last one on the hypervisor for that bridge

Also, I would add a test to verify that broadcast storm prevention rules are enforced correctly even after VM migration.

Thanks,
Salvatore


From: Sanjeev Neelarapu
Sent: 30 April 2012 07:25
To: Kevin Kluge; Salvatore Orlando; Chiradeep Vittal
Cc: #Cloud - Bonita
Subject: RE: Test cases for Open vSwitch Tunnel Manager

Hi,

Following is the link to test cases for Open vSwitch Tunnel Mananger:

http://wiki.cloudstack.org/display/QA/Open+vSwitch+Tunnel+Manager



Thanks,
Sanjeev

From: Kevin Kluge
Sent: Friday, April 27, 2012 11:36 PM
To: Sanjeev Neelarapu; Salvatore Orlando; Chiradeep Vittal
Cc: #Cloud - Bonita
Subject: RE: Test cases for Open vSwitch Tunnel Manager

Can we also post these on the wiki pls.

From: Sanjeev Neelarapu
Sent: Friday, April 27, 2012 6:00 AM
To: Salvatore Orlando; Chiradeep Vittal
Cc: #Cloud - Bonita
Subject: Test cases for Open vSwitch Tunnel Manager

Hi,

PFA initial test case document for "Open vSwitch Tunnel Manager".
Kindly review the test cases and send the feedback.

FS for OVS Tunnel Manager is @ http://confluence.cloudstack.org/display/RelOps/Open+vSwitch+tunnel+manager+for+Cloudstack


Thanks,
Sanjeev