Saul, The new repo and pending move was communicated in TSC meeting minutes on June 3rd. I subsequently gave StarlingX and documentation and test teams a heads up regarding the pending remote bootstrap instructions. I forgot about your sysadmin commit and consequently omitted to include you in that communication. This must be frustrating. Sorry :) If it helps, I can post a commit with sysadmin change in stx-ansible-playbooks repo and link it to your commit in root repo, make a dev build and run some tests. You will just need to rebase your commit in stx-config. Again, my apologies for the inconveniences this move has caused. Tee -----Original Message----- From: Saul Wold [mailto:sgw@linux.intel.com] Sent: June-14-19 5:22 PM To: starlingx-discuss@lists.starlingx.io Cc: Penney, Don Subject: Re: [Starlingx-discuss] Bootstrap playbook relocation This seems to have caused a problem with the wrsroot -> sysadmin changes, looks like I am going to have to redo a load of work now. I wish there had been some warning that this was going to happen so I could have weighed in on the effect to the sysadmin update. I think we should have completed the wrsroot-> sysadmin change first and then this move. Sau! On 6/14/19 1:05 PM, Ngo, Tee wrote:

Hi, The playbookconfig directory in StarlingX config repo, which contains the source code of the bootstrap playbook, has been relocated to stx-ansible-playbooks repo. Tee

_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss

_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss

On Fri, Jun 14, 2019 at 5:41 PM Ngo, Tee <Tee.Ngo@windriver.com> wrote:

Scott helped setting up the new repo. The git history import must have been missed as part of the setup. Does it mean we need to redo the repo creation?

Yes, the process for preserving history in a move like this is to create a new staging repo with the history using something like [0] then import that when creating it in Gerrit. At this point I think the only way to recover the history is to re-create the Gerrit repo. Any changes to the initial repo would need to be re-created. This is well into needing help from the OpenDev/Infra teams at this point... dt [0] https://github.com/dtroyer/home/blob/master/bin/split-repo.sh -- Dean Troyer dtroyer@gmail.com

On 2019-06-14 17:49:09 -0500 (-0500), Dean Troyer wrote: [...]

At this point I think the only way to recover the history is to re-create the Gerrit repo. Any changes to the initial repo would need to be re-created. This is well into needing help from the OpenDev/Infra teams at this point... [...]

Once you have an external, publicly cloneable repo somewhere from which I can pull the new refs, I'm happy push --force those over top of the old content. Or if you prefer, you can propose a temporary ACL change for that repo in openstack/project-config to grant some Gerrit group you're in access to do the same, and I'll be glad to review/approve it. -- Jeremy Stanley

I see the following: 2dc3228 Add .gitignore to ansible-stx-playbooks repo d3360f6 Populate stx-ansible-playbooks repo 6bef82a Initial zuul / TOX setup 4266117 Added .gitreview The second one on the list, d3360f6, was just moving the playbookconfig dir from stx-config, and population the centos_iso_image.inc and centos_pkg_dirs files for the build. I believe the first and second ones were also just copies of the files in stx-config, setting up the repo. -----Original Message----- From: Dean Troyer [mailto:dtroyer@gmail.com] Sent: Friday, June 14, 2019 8:16 PM To: Jeremy Stanley Cc: starlingx Subject: Re: [Starlingx-discuss] Bootstrap playbook relocation On Fri, Jun 14, 2019 at 6:44 PM Jeremy Stanley <fungi@yuggoth.org> wrote:

Once you have an external, publicly cloneable repo somewhere from which I can pull the new refs, I'm happy push --force those over top of the old content. Or if you prefer, you can propose a temporary ACL change for that repo in openstack/project-config to grant some Gerrit group you're in access to do the same, and I'll be glad to review/approve it.

Thanks Jeremy, I imagine we'll create it in starlingx-staging and push from there. It sounds like we'll see what Scott wants to do on Monday. There are 4 merged reviews in that project now, I know we'll have to re-create them. After a force-push will they still have anything interesting? I don't think it will be a big deal, just curious. dt -- Dean Troyer dtroyer@gmail.com _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss

Hi Dean, Jeremy, Scott, Could you do what are necessary to have the new repo up with git history. A few developers are waiting to update/add new playbooks to this repo. Thanks! Tee -----Original Message----- From: Scott Little [mailto:scott.little@windriver.com] Sent: June-17-19 11:07 AM To: starlingx-discuss@lists.starlingx.io Subject: Re: [Starlingx-discuss] Bootstrap playbook relocation Well, so as long as the Tee, and the authors of these reviews (if different), and any other interested parties are all aware of the forthcoming history rewrite, I won't object.  The interested community should be small for such a new repo. Scott On 2019-06-14 9:05 p.m., Penney, Don wrote:

I see the following: 2dc3228 Add .gitignore to ansible-stx-playbooks repo d3360f6 Populate stx-ansible-playbooks repo 6bef82a Initial zuul / TOX setup 4266117 Added .gitreview

The second one on the list, d3360f6, was just moving the playbookconfig dir from stx-config, and population the centos_iso_image.inc and centos_pkg_dirs files for the build. I believe the first and second ones were also just copies of the files in stx-config, setting up the repo.

-----Original Message----- From: Dean Troyer [mailto:dtroyer@gmail.com] Sent: Friday, June 14, 2019 8:16 PM To: Jeremy Stanley Cc: starlingx Subject: Re: [Starlingx-discuss] Bootstrap playbook relocation

On Fri, Jun 14, 2019 at 6:44 PM Jeremy Stanley <fungi@yuggoth.org> wrote:

...

Once you have an external, publicly cloneable repo somewhere from which I can pull the new refs, I'm happy push --force those over top of the old content. Or if you prefer, you can propose a temporary ACL change for that repo in openstack/project-config to grant some Gerrit group you're in access to do the same, and I'll be glad to review/approve it. Thanks Jeremy, I imagine we'll create it in starlingx-staging and push from there. It sounds like we'll see what Scott wants to do on Monday.

There are 4 merged reviews in that project now, I know we'll have to re-create them. After a force-push will they still have anything interesting? I don't think it will be a big deal, just curious.

dt

_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss

I'm ok with option b). Please proceed if there are no other objections. Scott On 2019-06-17 12:48 p.m., Dean Troyer wrote:

...

Well, so as long as the Tee, and the authors of these reviews (if different), and any other interested parties are all aware of the forthcoming history rewrite, I won't object. The interested community should be small for such a new repo. This is exactly why I didn't just ask fungi to force-push over the original repo. You guys get to make the call to a) just leave it all as-is and continue on; b) force-push my test or another similar exercise containing the original history; or c) start over from scratch. I would recommend b). The biggest concern I have about not fixing it is that I am planning to do more of these and want to set

On Mon, Jun 17, 2019 at 10:15 AM Scott Little <scott.little@windriver.com> wrote: the right precedent and process for doing extractions, especially when it will be code with a year's worth of history behind to rather than a few months. Also, I feel creating a new repo rather than force-pushing one is mostly-unnecessary work.

If you do choose b), please verify it to be correct. The playbookconfig directory compared exactly using diff -rw, the root dir needed a bit of work.

dt

On Fri, Jun 14, 2019 at 7:16 PM Dean Troyer <dtroyer@gmail.com> wrote:

Thanks Jeremy, I imagine we'll create it in starlingx-staging and push from there. It sounds like we'll see what Scott wants to do on Monday.

So I went ahead and did this to see how hard it would be...the result is in [0] and the steps I followed are in [1].

There are 4 merged reviews in that project now, I know we'll have to re-create them. After a force-push will they still have anything interesting? I don't think it will be a big deal, just curious.

In the end only commits 426611788310 (.gitreview) and 6bef82acb56d (https://review.opendev.org/#/c/664635/, tox.ini, .zuul.yaml) were used from the existing Gerrit repo. I added one additional commit to account for changes made to files in https://review.opendev.org/665437 after being copied from starlingx/config (they only appear as adds in 665437, the actual changes are not separately visible). [0] should be usable as an import/force-push to Gerrit to reset the ansible-playbooks repo if it looks good to everyone. dt [0] https://github.com/dtroyer/ansible-playbooks-staging-1 [1] https://gist.github.com/dtroyer/5d420b65d898019467dd2c9a03c15407 -- Dean Troyer dtroyer@gmail.com