[ovs-dev] [PATCH ovn] tests: Really fix requested-chassis localport test
numans at ovn.org
Mon Nov 29 18:25:24 UTC 2021
On Fri, Nov 26, 2021 at 7:52 AM Dumitru Ceara <dceara at redhat.com> wrote:
> On 11/8/21 10:38, Frode Nordahl wrote:
> > The first test case of adding localport without a
> > requested-chassis option is invalid as this would make the
> > chassis do claim/release thrashing leading to a unpredictable
> > test results.
> > Create the localport in NB DB prior to creating the ports on the
> > chassis. The step by step ovn-nbctl CLI creation of the localport
> > would in reality create a VIF which is then changed to be a
> > localport, this does not represent how the CMS would create the
> > localport.
> > Create a second port representing a VM to ensure the localport
> > flows are generated.
> > Use OVS_WAIT_UNTIL instead of AT_CHECK, which should avoid race
> > conditions where the check is ran before the flow tables are
> > populated.
> > Print output so that it is easier to investigate in the event of a
> > failure.
> > Fixes: ad77db239d9a ("controller: Make use of Port_Binding:requested_chassis.")
> > Signed-off-by: Frode Nordahl <frode.nordahl at canonical.com>
> > ---
> Hi Frode,
> I confirm this fixes the flaky test (or at least I haven't seen another
> failure with your patch applied while without I see them every now and
> Acked-by: Dumitru Ceara <dceara at redhat.com>
> dev mailing list
> dev at openvswitch.org
More information about the dev