Addons

Technical Overview

Updated: November 15, 2024

Dedicated Low-Level Fields

To provide the required data to Amazon S3 (expiration date and legal hold) and to ensure that records can’t be deleted when not eligible, some new low-level fields have been added at the storage level:

  • ecm:isRecord (boolean): true if the document is declared as a record.
  • ecm:isFlexibleRecord (boolean): true if the document is declared as a flexible record, enforced otherwise. This field cannot be true if ecm:isRecord is false.
  • ecm:retainUntil (datetime): Expiration date (end of retention period).
  • ecm:hasLegalHold (boolean): true if there is a legal hold applied to the record.
  • ecm:isUnderRetentionOrLegalHold (boolean): true if the record is under retention and/or legal hold.

If a legal hold is applied to a document or if a retention rule is applied to a document, then:

  • Modification of the main blob (file:content) and any other retained blob property is prevented: this includes modification through restoring an older version.
  • Deletion (including recursive delete) is prevented: this is enforced at low-level and this information is made available per-document through a security policy.
  • ecm:isUnderRetentionOrLegalHold is set to true.

This applies in all situations, no exception, including for Administrators.

In case of event-based retention, the record is put on a temporary hold waiting for the event to occur (see Retention flow section). In this case, ecm:retainUntil is "indeterminate" (by setting the field to 9999-01-01T00:00:00.000+00:00), so deletion is prevented as if it was a date far in the future (this is turned into a legal hold at the S3 storage level).

A copy of a document resets its record, retention and legal hold information. This includes creating a new version (that’s why versioning must be disabled to be compliant). These fields don't apply to proxies, as proxies are just a convenience to access content stored elsewhere, and are not themselves records.

Retention Flow

When a granted user applies a retention rule to a document, the document is declared as a record. The retention rule can create 2 types of record:

  • enforced : ecm:isRecord is true and ecm:isFlexibleRecord is false
  • flexible : ecm:isRecord is true and ecm:isFlexibleRecord is true

Enforced records have their retainable blobs copied to the dedicated S3 bucket where object lock will be applied according to retention policy of the attached rule.

The retainable blobs of Flexible records remains in the default bucket where no object lock is applied.

Then, depending on the retention rule type (immediate, metadata, or event-based retention), the expiration date is calculated and applied to both Nuxeo (Enforced and Flexible) and S3 object (Enforced only).

A background process checks when the event occurs and then starts the retention period (event-based retention).

Another background process checks the document expiration to change the status and make the record deletable. This process is run once a day (there may be a short delay for the record to become deletable).

retention-sequence-application
retention-sequence-application

When a granted user applies a legal hold to a record, the legal hold is applied at Nuxeo level and also at S3 level.

There are some specific use cases:

  • A legal hold is applied to a document which is not a record, the document is first declared as a record and moved to the compliance mode S3 bucket before applying the legal hold,

  • A legal hold is applied to a document which is a Flexible record, the document is turned into an enforced record and moved to the compliance mode S3 bucket before applying the legal hold,

  • There is an event-based retention rule applied to the document and the event hasn't occurred yet, a temporary hold is applied to the record. So, the legal hold is only applied at Nuxeo level (as it is already done on S3 level),

  • The document is already on legal hold, there is no action done except to log the legal hold action on the audit log.

retention-sequencing-legal-hold
retention-sequencing-legal-hold

Deletion Flow

Standard Deletion Flow

The standard Nuxeo deletion flow is based on the following principles:

  • Immediate deletion of the document content referenced in the Nuxeo database when a granted user requests deletion action,

  • As Nuxeo handles deduplication of the file content (when 2 Nuxeo documents contain the exact same file, the file is stored only once), the deletion of the files in the storage media is done as described in the Garbage Collector documentation:

    • The Garbage Collector checks the files stored on the storage media to be deleted are not referenced in Nuxeo Database as part of the document content,

    • If and only if a file is not referenced in any Nuxeo document (meaning all documents included the file have been deleted), the Garbage Collector process deletes the file in the storage media,

retention-standard-deletion
retention-standard-deletion

Alternative Record Deletion Flow

The Nuxeo Garbage Collector process is not available when some conditions are not met.

For the Retention Management addon, a specific deletion flow is available in the case the Garbage Collector process is not available:

  • The deduplication feature is turned off for enforced records content as it is not compliant with retention features (ex: it's not possible to handle a case where 2 records referenced the same file content with 2 different retention period and the legal constraint to delete it right after the retention period),

  • For compliance purpose, the deletion process is done synchronously when a granted user requests the deletion of an enforced record and when Nuxeo logs the deletion event.

retention-sequencing-deletion
retention-sequencing-deletion