Housekeeping (FREE SELF)
GitLab supports and automates housekeeping tasks in Git repositories to ensure that they can be served as efficiently as possible. Housekeeping tasks include:
- Compressing Git objects and revisions.
- Removing unreachable objects.
- Removing stale data like lock files.
- Maintaining data structures that improve performance.
- Updating object pools to improve object deduplication across forks.
WARNING: Do not manually execute Git commands to perform housekeeping in Git repositories that are controlled by GitLab. Doing so may lead to corrupt repositories and data loss.
Gitaly can perform housekeeping tasks in a Git repository in two ways:
- Eager housekeeping executes specific housekeeping tasks independent of the state a repository is in.
- Heuristical housekeeping executes housekeeping tasks based on a set of heuristics that determine what housekeeping tasks need to be executed based on the repository state.
The "eager" housekeeping strategy executes housekeeping tasks in a repository independent of the repository state. This is the default strategy as used by the manual trigger and the push-based trigger.
The eager housekeeping strategy is controlled by the GitLab application. Depending on the trigger that caused the housekeeping job to run, GitLab asks Gitaly to perform specific housekeeping tasks. Gitaly performs these tasks even if the repository is in an optimized state. As a result, this strategy can be inefficient in large repositories where performing the housekeeping tasks may be slow.
On self-managed GitLab, by default this feature is not available for the manual trigger and the push-based trigger.
To make it available, ask an administrator to enable the feature flag named
The heuristical (or "opportunistic") housekeeping strategy analyzes the
repository's state and executes housekeeping tasks only when it finds one or
more data structures are insufficiently optimized. This is the strategy used by
scheduled housekeeping. It can optionally be enabled
for the manual trigger and the push-based trigger
by enabling the
optimized_housekeeping feature flag.
Heuristical housekeeping uses the following information to decide on the tasks it needs to run:
- The number of loose and stale objects.
- The number of packfiles that contain already-compressed objects.
- The number of loose references.
- The presence of a commit-graph.
The decision whether any of the analyzed data structures need to be optimized is based on the size of the repository:
- Objects are repacked frequently the bigger the total size of all objects.
- References are repacked less frequently the more references there are in total.
Gitaly does this to offset the fact that optimizing those data structures takes more time the bigger they get. It is especially important in large monorepositories (which receive a lot of traffic) to avoid optimizing them too frequently.
Running housekeeping tasks
There are different ways in which GitLab runs housekeeping tasks:
- A project's administrator can manually trigger repository housekeeping tasks.
- GitLab can automatically schedule housekeeping tasks after a number of Git pushes.
- GitLab can schedule a job that runs housekeeping tasks for all repositories in a configurable time frame.
Administrators of repositories can manually trigger housekeeping tasks in a repository. In general this is not required as GitLab knows to automatically run housekeeping tasks. The manual trigger can be useful when either:
- A repository is known to require housekeeping.
- Automated push-based scheduling of housekeeping tasks has been disabled.
To trigger housekeeping tasks manually:
- On the top bar, select Main menu > Projects and find your project.
- On the left sidebar, select Settings > General.
- Expand Advanced.
- Select Run housekeeping.
This starts an asynchronous background worker for the project's repository. The
background worker executes
git gc, which performs a number of optimizations.
GitLab automatically runs repository housekeeping tasks after a configured number of pushes:
git gcruns a number of housekeeping tasks such as:
- Compressing Git objects to reduce disk space and increase performance.
- Removing unreachable objects that may have been created from changes to the repository, like force-overwriting branches.
- Runs an incremental repack, according to a configured period. This packs all loose objects into a new packfile and prunes the now-redundant loose objects.
- Runs a full repack, according to a configured period. This repacks all packfiles and loose objects into a single new packfile, and deletes the old now-redundant loose objects and packfiles. It also optionally creates bitmaps for the new packfile.
git pack-refscompresses references stored as loose files into a single file.
Configure push-based maintenance
You can change how often these tasks run when pushes occur, or you can turn them off entirely:
- On the top bar, select Main menu > Admin.
- On the left sidebar, select Settings > Repository.
- Expand Repository maintenance.
- In the Housekeeping section, configure the housekeeping options.
- Select Save changes.
The following housekeeping options are available:
- Enable automatic repository housekeeping: Regularly run housekeeping tasks. If you keep this setting disabled for a long time, Git repository access on your GitLab server becomes slower and your repositories use more disk space.
Incremental repack period: Number of Git pushes after which an incremental
git repackis run.
Full repack period: Number of Git pushes after which a full
git repackis run.
Git GC period: Number of Git pushes after which
git gcis run.
As an example, see the following scenario:
- Incremental repack period: 10.
- Full repack period: 50.
- Git GC period: 200.
pushes_since_gcvalue is 50, a
repack -A -l -d --pack-kept-objectsruns.
pushes_since_gcvalue is 200, a
Housekeeping also removes unreferenced LFS files
from your project on the same schedule as the
git gc operation, freeing up storage space for your
While GitLab automatically performs housekeeping tasks based on the number of pushes, it does not maintain repositories that don't receive any pushes at all. As a result, inactive repositories or repositories that are only getting read requests may not benefit from improvements in the repository housekeeping strategy.
Administrators can enable a background job that performs housekeeping in all repositories at a customizable interval to remedy this situation. This background job processes all repositories hosted by a Gitaly node in a random order and eagerly performs housekeeping tasks on them. The Gitaly node will stop processing repositories if it takes longer than the configured interval.
Configure scheduled housekeeping
Background maintenance of Git repositories is configured in Gitaly. By default, Gitaly performs background repository maintenance every day at 12:00 noon for a duration of 10 minutes.
You can change this default in Gitaly configuration. The following snippet
enables daily background repository maintenance starting at 23:00 for 1 hour
[daily_maintenance] start_hour = 23 start_minute = 00 duration = 1h storages = ["default"]
Use the following snippet to completely disable background repository maintenance:
[daily_maintenance] disabled = true
Object pool repositories
Object pool repositories are used by GitLab to deduplicate objects across forks of a repository. When creating the first fork, we:
- Create an object pool repository that contains all objects of the repository that is about to be forked.
- Link the repository to this new object pool via Git's altenates mechanism.
- Repack the repository so that it uses objects from the object pool. It thus can drop its own copy of the objects.
Any forks of this repository can now link against the object pool and thus only have to keep objects that diverge from the primary repository.
GitLab needs to perform special housekeeping operations in object pools:
- Gitaly cannot ever delete unreachable objects from object pools because they might be used by any of the forks that are connected to it.
- Gitaly must keep all objects reachable due to the same reason. Object pools thus maintain references to unreachable "dangling" objects so that they don't ever get deleted.
- GitLab must update object pools regularly to pull in new objects that have been added in the primary repository. Otherwise, an object pool will become increasingly inefficient at deduplicating objects.
These housekeeping operations are performed by the specialized
FetchIntoObjectPool RPC that handles all of these special tasks while also
executing the regular housekeeping tasks we execute for normal Git
Object pools are getting optimized automatically whenever the primary member is getting garbage collected. Therefore, the cadence can be configured using the same Git GC period in that project.
If you need to manually invoke the RPC from a Rails console,
you can call
project.pool_repository.object_pool.fetch. This is a potentially
long-running task, though Gitaly times out after about 8 hours.