[cinder][stable] branch freeze for ocata, pike
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted. stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted. [0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html [1] https://review.opendev.org/#/c/742523/
Hi Infra Team, While reviewing Andreas' patch [1], I have realized, that Cinder's stable/ocata and stable/pike branches were not deleted yet, however those branches were EOL'd already (see mail below). According to the process [2], since the EOL patches have merged already, if @Brian doesn't object, can you please delete - cinder stable/ocata - cinder stable/pike Thanks in advance, Előd [1] https://review.opendev.org/#/c/750887/ [2] https://docs.openstack.org/project-team-guide/stable-branches.html#end-of-li... On 2020. 07. 28. 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted.
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html [1] https://review.opendev.org/#/c/742523/
On 9/10/20 8:40 AM, Előd Illés wrote:
Hi Infra Team,
While reviewing Andreas' patch [1], I have realized, that Cinder's stable/ocata and stable/pike branches were not deleted yet, however those branches were EOL'd already (see mail below).
According to the process [2], since the EOL patches have merged already, if @Brian doesn't object, can you please delete
- cinder stable/ocata - cinder stable/pike
I have no objection, but I haven't pushed the infra team about the actual branch deletion because as far as I know, cinder is the first project to actually request removal, and I suspect all sorts of stuff will break. I suggest we wait until at least after RC-time to give us all one less thing to worry about. As far as avoiding breakage goes, I put up two patches to devstack so that it will check out the -eol tag of cinder/brick/cinderclient instead of the stable branch, but I suspect these only scratch the surface of what can be broken once the cinder project branches are deleted. https://review.opendev.org/#/c/742953/ https://review.opendev.org/#/c/742952/ Sean suggested in an earlier thread on this topic [0] that instead of deleting very old EM branches that some projects have EOL'd project-by-project, we should just delete them wholesale across openstack. That makes a lot of sense to me. [0] http://lists.openstack.org/pipermail/openstack-discuss/2020-May/015115.html cheers, brian
Thanks in advance,
Előd
[1] https://review.opendev.org/#/c/750887/ [2] https://docs.openstack.org/project-team-guide/stable-branches.html#end-of-li...
On 2020. 07. 28. 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted.
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html
On 2020. 09. 10. 15:30, Brian Rosmaita wrote:
On 9/10/20 8:40 AM, Előd Illés wrote:
Hi Infra Team,
While reviewing Andreas' patch [1], I have realized, that Cinder's stable/ocata and stable/pike branches were not deleted yet, however those branches were EOL'd already (see mail below).
According to the process [2], since the EOL patches have merged already, if @Brian doesn't object, can you please delete
- cinder stable/ocata - cinder stable/pike
I have no objection, but I haven't pushed the infra team about the actual branch deletion because as far as I know, cinder is the first project to actually request removal, and I suspect all sorts of stuff will break. I suggest we wait until at least after RC-time to give us all one less thing to worry about.
Sound good to me, thanks Brian!
As far as avoiding breakage goes, I put up two patches to devstack so that it will check out the -eol tag of cinder/brick/cinderclient instead of the stable branch, but I suspect these only scratch the surface of what can be broken once the cinder project branches are deleted.
https://review.opendev.org/#/c/742953/ https://review.opendev.org/#/c/742952/
Sean suggested in an earlier thread on this topic [0] that instead of deleting very old EM branches that some projects have EOL'd project-by-project, we should just delete them wholesale across openstack. That makes a lot of sense to me.
Ocata is quite abandoned nowadays so that makes sense. However, Pike has been active in the past months [3] more or less, so mass-deletion is not an option for Pike, I think. Thanks, Előd [3] https://review.opendev.org/#/q/branch:stable/pike+status:merged
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-May/015115.html
cheers, brian
Thanks in advance,
Előd
[1] https://review.opendev.org/#/c/750887/ [2] https://docs.openstack.org/project-team-guide/stable-branches.html#end-of-li...
On 2020. 07. 28. 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted.
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html
On 28.07.20 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted.
When do you plan to delete those branches? We have Zuul jobs that are broken, for example due to removal of devstack-plugin-zmq and we either should remove these from the branch or delete the branch. Currently Zuul complains about broken jobs. The two changes I talk about are: https://review.opendev.org/750887 https://review.opendev.org/750886 Andreas
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html
-- Andreas Jaeger aj@suse.com Twitter: jaegerandi SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, D 90409 Nürnberg (HRB 36809, AG Nürnberg) GF: Felix Imendörffer GPG fingerprint = EF18 1673 38C4 A372 86B1 E699 5294 24A3 FF91 2ACB
On Thu, Sep 10, 2020, at 5:42 AM, Andreas Jaeger wrote:
On 28.07.20 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted.
When do you plan to delete those branches? We have Zuul jobs that are broken, for example due to removal of devstack-plugin-zmq and we either should remove these from the branch or delete the branch. Currently Zuul complains about broken jobs.
The two changes I talk about are: https://review.opendev.org/750887 https://review.opendev.org/750886
I think we should go ahead and land those if we are waiting for a coordinated branch deletion. The zuul configs are branch specific and should be adjustable outside of normal backport procedures, particularly if they are causing problems like global zuul config errors. We've force merged some of these changes on stable branches in other projects if CI is generally unstable. Let us know if that is appropriate for this situation as well.
Andreas
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html
Just discussed with Clark and Fungi on IRC, that since the branches are already tagged (*-eol), merging the patches could cause some confusions. So it's easier to just wait until RC, as Brian suggested. Thanks, Előd On 2020. 09. 10. 18:49, Clark Boylan wrote:
On Thu, Sep 10, 2020, at 5:42 AM, Andreas Jaeger wrote:
On 28.07.20 16:23, Brian Rosmaita wrote:
tl;dr - do not approve any backports to stable/ocata or stable/pike in any Cinder project deliverable
stable/ocata has been tagged with ocata-eol in cinder, os-brick, python-cinderclient, and python-brick-cinderclient-ext. Nothing should be merged into stable/ocata in any of these repositories during the interim period before the branches are deleted. When do you plan to delete those branches? We have Zuul jobs that are broken, for example due to removal of devstack-plugin-zmq and we either should remove these from the branch or delete the branch. Currently Zuul complains about broken jobs.
The two changes I talk about are: https://review.opendev.org/750887 https://review.opendev.org/750886 I think we should go ahead and land those if we are waiting for a coordinated branch deletion. The zuul configs are branch specific and should be adjustable outside of normal backport procedures, particularly if they are causing problems like global zuul config errors. We've force merged some of these changes on stable branches in other projects if CI is generally unstable. Let us know if that is appropriate for this situation as well.
Andreas
stable/pike: the changes discussed in [0] have merged, and I've proposed the pike-eol tags [1]. Nothing should be merged into stable/pike in any of our code repositories from now until the branches are deleted.
[0] http://lists.openstack.org/pipermail/openstack-discuss/2020-July/016076.html
participants (4)
-
Andreas Jaeger
-
Brian Rosmaita
-
Clark Boylan
-
Előd Illés