Managing Disk Space

Posit Connect can consume significant disk space as users deploy content over time. This page describes the main sources of disk usage and the configuration settings available to manage them.

Most disk usage accumulates in the Connect data directory (Server.DataDir, which defaults to /var/lib/rstudio-connect). You can check the current usage with:

Terminal
du -sh /var/lib/rstudio-connect

Bundle retention

Each deployment creates a new bundle on disk. By default, all previous bundles are retained indefinitely, which can consume significant space.

To limit the number of bundles retained per content item, set Applications.BundleRetentionLimit:

/etc/rstudio-connect/rstudio-connect.gcfg
[Applications]
BundleRetentionLimit = 3

Older bundles are deleted periodically, which runs on the interval defined by Applications.BundleReapFrequency (default: 24h).

The limit counts only inactive bundles. The active bundle, and any bundle used by a report variant, are always retained in addition to the limit.

Publishers can also override this limit for an individual content item. This is useful if the content’s full history must be retained for business purposes, even when a server-wide limit is in effect, or conversely if some specific content bundles are known to be large and do not need to be retained.

Rendering output retention

Rendered content produces output each time it runs. Two settings control automatic cleanup of old renderings:

The cleanup frequency is controlled by Applications.RenderingSweepFrequency (default: 1h).

Python environment cleanup

Python content uses virtual environments stored on disk. Connect removes unused environments automatically. For details on how environment caching works, see Environment caching.

The cleanup interval is controlled by Applications.PythonEnvironmentReapFrequency (default: 24h). To disable cleanup, set Applications.PythonEnvironmentReaping to false.

Job history cleanup

Connect stores metadata and output for completed jobs on disk. Two settings control automatic cleanup:

A job is removed if either limit is exceeded.

Deleted content

When content is deleted, Connect normally moves its artifacts, including source bundle archives and directory trees, into {Server.DataDir}/trash and removes them in the background. If an artifact cannot be moved into trash, Connect removes it synchronously. The content is unavailable immediately, but its disk space might not be reclaimed until a later cleanup pass.

The cleanup interval is controlled by Applications.TrashReapFrequency (default: 5m). The optional Applications.TrashRetention setting controls how long retired artifacts remain on disk before cleanup (default: 0, which permits deletion on the next pass). Entries that cannot be removed remain in trash and are retried during later passes.

Runtime caches

Package caches are not automatically deleted when content is removed. When you delete a runtime cache, Connect moves it into {Server.DataDir}/trash and removes it in the background, the same as deleted content. The Applications.TrashReapFrequency and Applications.TrashRetention settings also apply to deleted runtime caches. See the Runtime Caches documentation for details on viewing and deleting these caches.

Applying changes

After updating the configuration file, restart Connect for the changes to take effect:

Terminal
sudo systemctl restart rstudio-connect

Summary of space-saving settings

The following table summarizes all settings that help manage disk space:

Setting Default Description
Jobs.MaxCompleted 1000 Maximum completed jobs per application
Jobs.OldestCompleted 30d Maximum age of completed jobs
Applications.RenderingSweepLimit 30 Maximum renderings per application
Applications.RenderingSweepAge 30d Maximum age of renderings
Applications.RenderingSweepFrequency 1h Rendering cleanup frequency
Applications.BundleRetentionLimit 0 Maximum bundles per application (0 = unlimited)
Applications.BundleReapFrequency 24h Bundle cleanup frequency
Applications.TrashReapFrequency 5m Retired artifact cleanup frequency
Applications.TrashRetention 0 Minimum time retired artifacts remain on disk
Applications.PythonEnvironmentReaping true Enable Python environment cleanup
Applications.PythonEnvironmentReapFrequency 24h Python environment cleanup frequency