[ovs-dev] [PATCH] xenserver: Support fail_mode override at network level.

Ben Pfaff blp at nicira.com
Thu Jan 27 04:39:02 UTC 2011


It's because networks and bridges are not one-to-one.  In past experience
this mismatch has caused many surprises.  I can't see a reason why it wouldn't
work here, but I've learned to be really pessimistic about this code.

On Wed, Jan 26, 2011 at 8:34 PM, Ethan Jackson <ethan at nicira.com> wrote:
> Make check passes as expected.  I did not test it's behavior with
> respect to vlan's and bonds.  I can go ahead and test that now,  Is
> there a particular reason that you would expect it to fail in that
> case?  What worries you about the configure datapath change?
>
> Ethan
>
> On Wed, Jan 26, 2011 at 7:53 PM, Ben Pfaff <blp at nicira.com> wrote:
>> On Wed, Jan 26, 2011 at 6:18 PM, Ethan Jackson <ethan at nicira.com> wrote:
>>> This patch allows the Pool object's vswitch-controller-fail-mode
>>> setting to be overridden on by Network object.
>>>
>>> Signed-off-by: Ethan Jackson <ethan at nicira.com>
>>
>> I'm a little nervous about the change to configure_datapath().  You said
>> off-list that you tested this.  Did you test that it behaves OK
>> when there are VLANs and bonds involved?  Did you check that the unit
>> tests still pass?
>>
>> Otherwise it looks OK to me.
>>
>




More information about the dev mailing list