GitHub import: Run housekeeping after initial importsh-git-gc-after-initial-fetch

After an initial fetch, the repository will have many loose objects
and refs. Running a `git gc` will compact refs into `packed-refs` and
objects into `.pack` files, which should make importing pull
requests faster.

Part of https://gitlab.com/gitlab-org/gitlab-ce/issues/59477

3 files changed, 15 insertions, 0 deletions

diff --git a/changelogs/unreleased/sh-git-gc-after-initial-fetch.yml b/changelogs/unreleased/sh-git-gc-after-initial-fetch.yml
new file mode 100644
index 00000000000..867d7e6b9df
--- /dev/null
+++ b/changelogs/unreleased/sh-git-gc-after-initial-fetch.yml
@@ -0,0 +1,5 @@
+---
+title: 'GitHub import: Run housekeeping after initial import'
+merge_request: 26600
+author:
+type: performance
diff --git a/lib/gitlab/github_import/importer/repository_importer.rb b/lib/gitlab/github_import/importer/repository_importer.rb
index 6d48c6a15b4..6aad7955415 100644
--- a/lib/gitlab/github_import/importer/repository_importer.rb
+++ b/lib/gitlab/github_import/importer/repository_importer.rb
@@ -54,6 +54,11 @@ module Gitlab
           project.repository.fetch_as_mirror(project.import_url, refmap: refmap, forced: true, remote_name: 'github')
 
           project.change_head(default_branch) if default_branch
+
+          # The initial fetch can bring in lots of loose refs and objects.
+          # Running a `git gc` will make importing pull requests faster.
+          Projects::HousekeepingService.new(project, :gc).execute
+
           true
         rescue Gitlab::Git::Repository::NoRepository, Gitlab::Shell::Error => e
           fail_import("Failed to import the repository: #{e.message}")
diff --git a/spec/lib/gitlab/github_import/importer/repository_importer_spec.rb b/spec/lib/gitlab/github_import/importer/repository_importer_spec.rb
index 41810a8ec03..705df1f4fe7 100644
--- a/spec/lib/gitlab/github_import/importer/repository_importer_spec.rb
+++ b/spec/lib/gitlab/github_import/importer/repository_importer_spec.rb
@@ -197,6 +197,11 @@ describe Gitlab::GithubImport::Importer::RepositoryImporter do
         .to receive(:fetch_as_mirror)
         .with(project.import_url, refmap: Gitlab::GithubImport.refmap, forced: true, remote_name: 'github')
 
+      service = double
+      expect(Projects::HousekeepingService)
+        .to receive(:new).with(project, :gc).and_return(service)
+      expect(service).to receive(:execute)
+
       expect(importer.import_repository).to eq(true)
     end