- Remove Jenkins pipeline from zuul/layout.yaml file in integration/config.git.
- Empty repository, leaving only a note in the README file. - https://gerrit.wikimedia.org/r/#/c/operations/puppet/cdh/+/518972/
- Archive any documentation page on mediawiki.org.
- Archive any Phabricator project for the extension.
- Edit any Phabricator project description, and add link to this ticket.
- Mark the repository read-only in Gerrit, and prefix description with [ARCHIVED] .
- Archive the repo mirror in Phabricator Diffusion.
- Archive the repo mirror on GitHub.com.
Description
Details
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | elukey | T226466 Move the puppet cdh and zookeeper submodules into operations/puppet | |||
Resolved | elukey | T226474 Archive cdh puppet submodule |
Event Timeline
Change 518913 had a related patch set uploaded (by Elukey; owner: Elukey):
[integration/config@master] Archive the cdh puppet submodule
Change 518913 merged by jenkins-bot:
[integration/config@master] Archive the cdh/cdh4 puppet submodule
Mentioned in SAL (#wikimedia-releng) [2019-06-25T11:42:14Z] <hashar> Deleting Jenkins jobs puppet-cdh-rake-docker and puppet-cdh4-rake-docker # T226474
Why do we want to delete everything? Can't we just leave it up with a note in the README me that active development has moved into ops/puppet?
From what I know this is the procedure, I followed the same thing the last time for the other modules when I merged them. History is still available, so if somebody wants the code it is easy to get the last commit state or check operations/puppet..
I guess, just kind of a shame since now google will stop indexing and returning it in search results. It'd be nice if people could actually find it.
I don't find the documentation that states the procedure in the description, the last time @Krinkle showed it to me, maybe there are some caveats to get around it.
Indexing will be done in the operations/puppet repo though, that will be the most up to date version of the module no?
Yeah, but it will be lost to non WMFers, since it is buried deep in a huge
repo. Oh well I guess :(
Change 520254 had a related patch set uploaded (by Krinkle; owner: Krinkle):
[operations/puppet/cdh@master] readme: Add link to new location
Change 520254 merged by Krinkle:
[operations/puppet/cdh@master] readme: Add link to new location
Mentioned in SAL (#wikimedia-operations) [2019-07-02T15:20:18Z] <Krinkle> Set repo back from active to read-only https://gerrit.wikimedia.org/r/#/admin/projects/operations/puppet/cdh (T226474))
On the GitHub mirror, I've closed outstanding pull requests on the mirror, and set the the "Archived" (read-only) flag on. This means it is still searchable and indexed, but no new pulls can be opened, and no new commits can be mirrored.
Yes archive it in Diffusion and we will also just delete the Github mirror.
Just a note, it is possible to merge the repository into operations/puppet.git while keeping the full history. Though it is probably too late now, the process would look like:
Prepare the repo to be merged by moving all files:
git clone operations/puppet/cdh /tmp/cdh mkdir -p modules/cdh git mv * modules/cdh git commit -a -m 'Move everything to prepare for repo merge'
Then on the destination repo fetch the above:
cd projects/operations/puppet git remote add tobemerged/tmp/cdh git fetch tobemerged // verify head of the fetched repository match commit from above: git log --oneline FETCH_HEAD // """Move everything to prepare for repo merge"""
And do the merge
mkdir -p modules/cdh git merge FETCH_HEAD
Exact commands to be determined, but I have done that several time on multiple repositories. Maybe that can help if need be.
IIUC Timo suggested to leave the github mirror there as archived, but I don't have any strong opinion on this. Will try to archive it in Diffusion then :)
Yes, I mostly changed our practice of deleting mirrors to archiving them (which means it's fully ready-only, including no pull requests or other sources of notifications, and shows a nice banner on top at GitHub). This way doesn't break urls or discovery paths that people might have placed in blogs, code comments, toots, books, etc.
When we first started this, GitHub didn't have an archival feature.