We were seeing a number of error messages when attempting to create a keep-around ref:

1. Failed to create locked file `refs/keep-around/XYZ`: File exists
2. Failed to write reference `refs/keep-around/XYZ`: a reference with that name already exists.

I'm not sure how these happen, but I suspect when multiple workers attempt to write the same
file we may have an issue. The force parameter should help ensure the file gets created,
as well as the rescues to prevent 500 Errors.

Rugged/libgit2 unfortunately do not allow you to delete or re-create a reference that has
been corrupted, even with the force parameter.

Closes #20109

Signed-off-by: NRémy Coutable <remy@rymai.me>

Merge branch '20107-update-documentation-according-to-the-new-multiple-repository-storage-paths-feature' into 'master'

Update documentation according to the new multiple git mount points feature

Closes #20107

See merge request !5411

Refactor Slack service documentation



See merge request !5421

[ci skip]

Move admin application settings to own path

Move existing documentation to its own `user/admin_area/settings` path.

See merge request !5420

Refactor protected branches documentation

Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/19832

See merge request !5419

[ci skip]

Simpler two queries than one JOIN with subquery

This is a follow up from !5347

Originally it was:

``` ruby
pipeline = pipelines.latest_successful_for(ref)
builds.where(pipeline: pipeline).latest.with_artifacts
```

However MySQL would complain that we can't use `IN` against a subquery which has `LIMIT`. Using `INNER JOIN` would be a workaround, however, doing that is too complicated in current version of Rails.

So let's just use two queries in this case.

Closes #14419

See merge request !5388

Explain CI_PROJECT_NAMESPACE better

## What does this MR do?

Explain CI_PROJECT_NAMESPACE better by adding "The project namespace (username or groupname) that is currently being built".

## Are there points in the code the reviewer needs to double check?

Are `username` and `groupname` known words? We use it [here](http://docs.gitlab.com/ee/api/namespaces.html).

## Why was this MR needed?

Because `namespace` isn't a well known term.

## What are the relevant issue numbers?

## Screenshots (if relevant)


See merge request !5407

Add a spec for #20079

- This MR adds a spec for the fix for #20079 that was implemented in !5400.
- The spec is going in separately, just so the fix could go in as soon as possible.

See merge request !5414

The issue was fixed in 2ba5e625. The spec is going in separately just
so the fix could go in as soon as possible.

Fix a false positive where similar un-nested storage paths were reported as nested

Closes #19572

See merge request !5145

Merge branch '17073-tagscontroller-index-is-terrible-response-time-goes-up-to-5-seconds' into 'master'

Improve performance of tags controller

See merge request !5375

Storage path are not created until `TestEnv.init`, so we must skip
their validation on initialization.

Change nav link snippet controller

Removes active Snippets nav link state for project-specific snippets

Closes #20046 

See merge request !5410

Avoid data-integrity issue when cleaning up repository archive cache

## What does this MR do?
Sets the default value for `repository_downloads_path` if someone has it configured incorrectly, and it points to the path where repositories are stored. It's also replace invocation of `find` with Ruby code that matches old cached files in a better, and safe way to avoid data-integrity issues.

## Why was this MR needed?
The `repository_downloads_path` is used by the `RepositoryArchiveCacheWorker` to remove outdated repository archives, if it points to the wrong directory can cause some data-integrity issue.

## What are the relevant issue numbers?
Closes #14222

See merge request !5285

Reduce min width of pipeline table

## What does this MR do?
Reduce min width of pipeline table

## Why was this MR needed?
The current min-width is based on the heavily populated pipelines page main table. This pipeline table doesn't have nearly as much information so can be collapsed down a lot further

![table](/uploads/895d2fb2f01383a841f973e918199f55/table.gif)   
cc @markpundsack 


See merge request !5409

Retrieve rendered HTML from cache in one request

## What does this MR do?
It consolidates requests made to cache store to check for cached HTML of multiple Markdown texts (for reference extracting purposes) into one request.

## Are there points in the code the reviewer needs to double check?
N/A

## Why was this MR needed?
To improve the performance of `Participable#raw_participants`

## What are the relevant issue numbers?
#19985 

## Screenshots (if relevant)
N/A

## Does this MR meet the acceptance criteria?

- [x] [CHANGELOG](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CHANGELOG) entry added
- ~~[ ] [Documentation created/updated](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/development/doc_styleguide.md)~~
- ~~[ ] API support added~~
- ~~Tests~~
  - ~~[ ] Added for this feature/bug~~
  - ~~[ ] All builds are passing~~
- [x] Conform by the [style guides](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CONTRIBUTING.md#style-guides)
- [x] Branch has no merge conflicts with `master` (if you do - rebase it please)
- [x] [Squashed related commits together](https://git-scm.com/book/en/Git-Tools-Rewriting-History#Squashing-Commits)

See merge request !5406

Refactor Issues::BulkUpdateService spec

- Create fewer Issue objects; 2 is as good as 5 for these cases. This
  gives us a nice little speed improvement.
- Don't `describe` Symbols.
- Simplify object creation.
- Lessen "mystery guest" antipattern

See merge request !5380

See #19985

Add Discussion model to represent MR/diff discussion

Related to https://gitlab.com/gitlab-org/gitlab-ce/issues/10325.

See merge request !5376

UI bug fixes

Closes #20074   
Closes #20094 

Regression UI bug fixes
* Fix broken status icon on project page
* Center tooltip on empty state in pipeline
* Reduce max-width of commit title in pipeline so more fits in table
* Fixes firefox rendering (icons getting cut off)
* Fixes Safari rendering of SHA icon

See merge request !5383

Bump vmstat version to fix issues reporting on FreeBSD

Minor changes from 2.1.0 to 2.1.1:

https://github.com/threez/ruby-vmstat/commits/master

See merge request !5405

Don't drop in DropAndReaddHasExternalWikiInProjects

## What does this MR do?

This MR fixes the migration in question so it doesn't first drop and then re-add a column.

## Are there points in the code the reviewer needs to double check?

Not really.

## Why was this MR needed?

The old migration could lead to application errors as the column would temporarily not exist between the drop column and add column statements.

## What are the relevant issue numbers?

https://gitlab.com/gitlab-com/infrastructure/issues/239

## Does this MR meet the acceptance criteria?

- [x] ~~[CHANGELOG](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CHANGELOG) entry added~~
- [x] ~~[Documentation created/updated](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/development/doc_styleguide.md)~~
- [x] Conform by the [style guides](https://gitlab.com/gitlab-org/gitlab-ce/blob/master/CONTRIBUTING.md#style-guides)
- [ ] Branch has no merge conflicts with `master` (if you do - rebase it please)
- [x] [Squashed related commits together](https://git-scm.com/book/en/Git-Tools-Rewriting-History#Squashing-Commits)

See merge request !5402

Add allow_failure CI documentation

## What does this MR do?

Adds documentation for the `allow_failure` keyword for the CI.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/17669

See merge request !5398

Dropping a column and then re-adding it can lead to the application
throwing errors as the column may temporarily not exist. To work around
this we'll reset the various project rows in batches _without_ removing
any columns.

Don't use `params[:id]` while building `markdown_preview_path`.

- Closes #20079

See merge request !5400