This documentation explains what options are available to clean up your orphaned content, which one(s) you should pick based on your situation and how to apply them.
How Nuxeo Cleans Up Content
Nuxeo Server is designed to give you complete control over your content and how you dispose of it.
For example, you may need to keep logs to prove compliance even if the document was deleted, or you may want to keep completed workflows around in order to build analytics and gather insights. To cover for such use cases, Nuxeo Server offers specific ways to dispose of orphaned content, and you need to leverage them explicitly.
Read on to discover what options are available to you, when, how and why you should leverage them.
Available Options
Cleaning Up Orphaned Versions for Deleted Documents
When deleting permanently a document, all its versions are deleted automatically. There are some scenarios where a version will remain however.
A version stays referenced and therefore is not removed if:
- Any proxy points to a version in the version history of any live document
- In the case of a tree snapshot if there is a snapshot containing a version in the version history of any live document.
- In some specific scenarios when deleting documents recursively (e.g., some versions of a child document may remain if you delete a parent document)
If you are confident that older versions of deleted documents are not in use anymore, Nuxeo Server offers a dedicated REST API endpoint to remove them.
This endpoint leverages a bulk action, meaning that it is scalable for any repository size and can be easily monitored, even in a context where it could end up being a long-running action.
Cleaning Up Orphaned Binaries
Nuxeo Server deduplicates content in order to optimize storage costs. This means that multiple documents may reference the same file, and yet it will be stored only once.
When deleting permanently a document, other documents may still reference some or all of the files it holds. If not, then these files are considered orphaned.
If you started using Nuxeo after LTS 2021 HF35, orphaned files are cleaned up automatically through a background task. If you chose to disable this feature, an on demand option exists as well.
If you started using Nuxeo before LTS 2021 HF35, then you may need to run some migration steps to gain access to this feature. This will prevent new orphaned files to be created. Existing orphaned files can be cleaned up using the on demand cleanup option.
You can learn how to use this option in the Garbage Collecting Orphaned Blobs documentation.
Cleaning Up Completed Workflows
Workflows are stored as documents in Nuxeo Server. Each time a workflow is started, new documents are created for each step in the workflow. This means that if you leverage workflows heavily, your document count will grow along.
To mitigate this, workflows in state "done" or "canceled" are deleted automatically on a daily basis along with all the steps they contained. And since LTS 2021 HF35, this automatic removal uses a bulk action to make it even more scalable.
In addition to this mechanism, it is possible to do further cleanup with the following actions:
By setting the nuxeo.routing.cleanup.workflow.instances.orphan nuxeo.conf property to true (false by default), the DocumentRoutingWorkflowInstancesCleanup listener will also remove orphan workflows daily.
A new management rest API endpoint is also available to perform a workflow garbage collection on demand. See the workflows endpoint documentation. Reasons to use it include:
- You disabled the automatic workflow cleanup using configuration in order to build reports or gather data on your workflows
- You did not enable the
nuxeo.routing.cleanup.workflow.instances.orphannuxeo.conf property yet - Before LTS 2021 HF35, workflows were not cleaned up in some specific scenarios
Cleaning Up Audit Logs
Audit logs keep track of all the activity for documents handled in Nuxeo Server. You may need it for security audits or to prove compliance for example. Therefore, it is kept forever until you explicitly choose to delete it, even for deleted documents.
You can learn how to use this option in our dedicated Purging Audit Logs documentation.