Hi Ales

On Wed, Jan 21, 2026 at 9:15 AM Ales Musil <[email protected]> wrote:

>
>
> On Wed, Jan 21, 2026 at 9:12 AM Xavier Simonart <[email protected]>
> wrote:
>
>> Hi Ales
>>
>> Thanks for looking into this.
>>
>> On Wed, Jan 21, 2026 at 8:40 AM Ales Musil <[email protected]> wrote:
>>
>>>
>>>
>>> On Mon, Jan 19, 2026 at 2:15 PM Xavier Simonart via dev <
>>> [email protected]> wrote:
>>>
>>>> This will tag tests for which upgrade tests would fail.
>>>> This is a pre-requisite for upgrade test patch [1].
>>>> When upgrade tests (from [1]) are run, tests tagged with
>>>> TAG_TEST_NOT_UPGRADABLE will be skipped.
>>>>
>>>> The "NOT_UPGRADABLE" test would fail since patch [2] added a
>>>> new table, which is unexpected by the old test.
>>>>
>>>> [1] ci: Run system tests in upgrade scenario
>>>> [2] controller: Learn EVPN workload FDB from the ingress traffic
>>>>
>>>> Signed-off-by: Xavier Simonart <[email protected]>
>>>> ---
>>>>
>>>
>>> Hi Xavier,
>>>
>>> thank you for the patch. I think this is a little wrong way of doing
>>> this.
>>> We should have this patch on main first and then backport it.
>>> After that we can proceed with the upgrade testing, would you agree?
>>>
>> I am not sure.
>>
>> The upgrade test fails to upgrade to main. So, the tag itself
>> (TAG_TEST_NOT_UPGRADABLE
>> in system-ovn.at) should only be set in this branch. It should not be
>> set on main.
>> Of course, we could have this patch on main, backport it to this branch
>> only, and then the upgrade patch itself, in main, would also remove the tag
>> from the test.
>> WDYT?
>> By the way, the fact that this patch is not (yet) in github branch-25.09
>> is the reason why the upgrade test failed (
>> https://github.com/ovsrobot/ovn/actions/runs/21139194130)
>>
>
> I was referring mainly to the addition of the tag itself, but having a
> custom backport of the tag as it's this patch here would solve that. So
> what I'm thinking of is a patch that adds the tag on main, then this custom
> backport that also adds it to the test on 25.09, that will refer to
> the patch on main. Does that make more sense?
>
Yes, I think it makes sense . This way this patch (with macro change and no
system-test change) has a reference on main, can be custom backported (with
macro change + system-test change) and we do not set/unset the tag in
system-test on main which was kind of overkill.
Thanks
Xavier

>
>
>> Thanks
>> Xavier
>>
>>
>>>
>>>>  tests/ovn-macros.at | 6 ++++++
>>>>  tests/system-ovn.at | 1 +
>>>>  2 files changed, 7 insertions(+)
>>>>
>>>> diff --git a/tests/ovn-macros.at b/tests/ovn-macros.at
>>>> index 94a0d61ef..0b8599d31 100644
>>>> --- a/tests/ovn-macros.at
>>>> +++ b/tests/ovn-macros.at
>>>> @@ -1493,6 +1493,12 @@ m4_define([TAG_UNSTABLE], [
>>>>      AT_SKIP_IF([test X"$SKIP_UNSTABLE" = Xyes])
>>>>  ])
>>>>
>>>> +# TAG_TEST_NOT_UPGRADABLE tag indicates that the test would fail
>>>> +# "upgrade" test (i.e. running old ovn-northd and new ovn-controller)
>>>> +m4_define([TAG_TEST_NOT_UPGRADABLE], [
>>>> +    AT_SKIP_IF([test X"$UPGRADE_TEST" = Xyes])
>>>> +])
>>>> +
>>>>  m4_define([OVN_CHECK_SCAPY_EDNS_CLIENT_SUBNET_SUPPORT],
>>>>  [
>>>>      AT_SKIP_IF([test $HAVE_SCAPY = no])
>>>> diff --git a/tests/system-ovn.at b/tests/system-ovn.at
>>>> index 6aca7b7df..efe6ff545 100644
>>>> --- a/tests/system-ovn.at
>>>> +++ b/tests/system-ovn.at
>>>> @@ -17500,6 +17500,7 @@ AT_CLEANUP
>>>>  OVN_FOR_EACH_NORTHD([
>>>>  AT_SETUP([dynamic-routing - EVPN])
>>>>  AT_KEYWORDS([dynamic-routing])
>>>> +TAG_TEST_NOT_UPGRADABLE
>>>>
>>>>  CHECK_VRF()
>>>>  CHECK_CONNTRACK()
>>>> --
>>>> 2.47.1
>>>>
>>>> _______________________________________________
>>>> dev mailing list
>>>> [email protected]
>>>> https://mail.openvswitch.org/mailman/listinfo/ovs-dev
>>>>
>>>>
>>> Regards,
>>> Ales
>>>
>>
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to