On Tue, Jun 14, 2022 at 9:04 PM Jeremy Stanley <fungi@yuggoth.org> wrote:

On 2022-06-14 20:40:30 +0900 (+0900), Takashi Kajinami wrote: [...]

...

In the past, we abandoned backport without checking this and this resulted in broken commit id in some backports. (stable/train backports point to the commit id of ussuri backport, which was never merged). [...]

Not sure if it's too heavy-handed, but one workaround would be to add a Depends-On footer in the commit message of each backport referencing the change for the branch ahead of it. Then, Zuul would only ever allow the backport to merge if the same change on all newer branches had already merged successfully.

Yeah I think that would be a solution though it requires relatively big overhead, because we need to manually add that Depends-on line. I sometimes hope zuul can block backports depending on the "cherry picked from commit ..." line but it might be worth discussing as a potential enhancement with the zuul devs.

-- Jeremy Stanley

On Tue, Jun 14, 2022, at 5:56 AM, Jeremy Stanley wrote:

On 2022-06-14 21:22:37 +0900 (+0900), Takashi Kajinami wrote: [...]

...

I sometimes hope zuul can block backports depending on the "cherry picked from commit ..." line but it might be worth discussing as a potential enhancement with the zuul devs.

Definitely worth further discussion, but I suspect the possibility for cherry-picks from other repositories could make that challenging to differentiate.

I agree. I'm not sure this makes sense as a built in Zuul feature. However, you can always implement a backport sanity check job that fails if certain criteria are not met.

-- Jeremy Stanley

Attachments: * signature.asc

I know we just discussed this in IRC but adding a note here for anyone else who may be interested. In the end we have 2 problematic ones [1] and [2] - i.e. the train cherrypick merged for those and it references the victoria commit from the cherrypicked from in the commit message. I realise this isn't ideal but at least it is better than last time - maybe we'll have 0 on the next one ;) thanks for raising it Takashi regards, marios [1] https://review.opendev.org/c/openstack/tripleo-common/+/837804 [2] https://review.opendev.org/c/openstack/tripleo-heat-templates/+/830281 On Tue, Jun 14, 2022 at 2:40 PM Takashi Kajinami <tkajinam@redhat.com> wrote:

I'd like to remind you that we should check whether that patch has been backported to any older branches (ussuri, train) we abandon any victoria backport.

In the past, we abandoned backport without checking this and this resulted in broken commit id in some backports. (stable/train backports point to the commit id of ussuri backport, which was never merged).

If the patch is already merged in ussuri or train then IMO we should try landing the victoria backport. If not, then please make sure you update the stable/train backport to remove any reference to the victoria backport.

On Tue, Jun 14, 2022 at 7:41 PM Marios Andreou <marios@redhat.com> wrote:

...

Hello o/

as proposed at [1][2] we are moving ahead with transitioning stable/victoria tripleo repos to End Of Life.

In order to do this we need to have no open patches against stable/victoria.

So please stop posting new patches for tripleo stable/victoria and if you have open patches please merge or abandon these by next Tuesday 21st June (for reference, list of repos with currently open patches at [3]).

No problem if someone needs a few more days we can wait just please reach out here or on irc. If there are no such requests then I will abandon all open reviews on 21st and respin [2] to pick up any new commits between now and then.

thanks for your help,

marios

[1] http://lists.openstack.org/pipermail/openstack-discuss/2022-June/028829.html [2] https://review.opendev.org/c/openstack/releases/+/845148 [3] (list of repos with open reviews): * https://review.opendev.org/q/project:openstack%252Fos-net-config+status:open... * https://review.opendev.org/q/project:openstack%252Fpython-tripleoclient+stat... * https://review.opendev.org/q/project:openstack%252Ftripleo-ansible+status:op... * https://review.opendev.org/q/project:openstack%252Ftripleo-common+status:ope... * https://review.opendev.org/q/project:openstack%252Ftripleo-heat-templates+st... * https://review.opendev.org/q/project:openstack%252Ftripleo-image-elements+st... * https://review.opendev.org/q/project:openstack%252Ftripleo-puppet-elements+s... * https://review.opendev.org/q/project:openstack%252Ftripleo-validations+statu...