Document Operations¶
Use these checks when documents cannot be read, edited online, or found in search. Domain administrators investigate records and access; host administrators inspect shared storage, service configuration, and workers. Start with read-only checks and use a test document before changing shared services.
Storage and Recovery Scope¶
Document metadata, versions, folder associations, ACLs, profiles, and file records depend on application databases. File content can be read from SQL storage, the configured archive database, or external file repositories. Repositories can use Local Directory or Google Cloud Storage; enabled repositories receive replicated files.
See Database Settings for repository configuration and Get Status counts. A file-service picker for external uploads is a separate capability from a host file repository that stores replicated content.
Before a storage change or restore, inventory:
- system and repository databases containing metadata, associations, profiles, and file records;
- the archive database when configured;
- local/shared repository paths and cloud objects needed to retrieve file content;
- host configuration, integrated-service access, and encryption password when encrypted files are used;
- the full-text service configuration and index recovery/rebuild procedure.
Keep these recovery points consistent. File copies alone cannot reconstruct document titles, access rules, and version relationships; restoring only metadata may leave references to missing content. Replication is not protection against an incorrect application change or deletion.
Agree the write-quiescence window, database/provider-specific backup procedure, and rollback point with the deployment owner. Restore into an isolated environment first. Verify a recent and an older version, all attached files, profile fields, folder placement, and allowed/denied accounts. Confirm encrypted files can be read and rebuild or validate search indexes after the records and content are available. Do not reconnect a restored test environment to live queues or outgoing integrations unintentionally.
Preview and Online Editing¶
Preview capabilities depend on file format and the configured viewer. Online editing uses a WOPI-compatible editor; review Editor Services, its enabled service and Discovery Url. Do not use a working preview as evidence that editing is configured.
For an editing failure, first check document access and whether the same file downloads. Then check editor discovery, service reachability from the relevant hosts, and the specific error. After a configuration change, reopen the test file and complete a save/read-back cycle. Record the previous settings so they can be restored if the test fails.
Search and Background Jobs¶
Document saves queue full-text work. The document index includes title, attached-file text, and profile XML; file-text extraction depends on the extraction service and file format. Search results can lag behind a successful document save.
- Confirm the version exists and opens from the folder with the intended reader account.
- Check the document ACL and whether the version is published, archived, or deleted. A published folder list and a search index are different views; do not assume their visibility rules are identical.
- Review document full-text jobs and failures in Background Jobs. The document maintenance task requires the
Workflow.Document.FullTextworker; file repository replication usesWorkflow.File.Replication. - Review Workers, the full-text service health, and Tika Service Url when file extraction fails.
- Resolve the underlying service/job failure before retrying an index operation. Reindexing can affect load and search availability; use the environment's approved operational procedure and scope it to the affected collection.
- Search again as the affected reader and a denied reader. Confirm both expected results and access filtering.
Marking a version Deleted removes it from search and prevents the document indexer from adding that deleted version again. Archiving is a different state; do not use it as a substitute for deleting a search result or revoking access.
Audit and Document History¶
Use Audit Log Search to investigate document creation, update, view, deletion, and folder events. Correlate the time, identity, document version ID, and branch ID. Delete audit data can distinguish permanent removal from marking a version deleted.
The document's History tab lists published versions; it is not an audit trail of every metadata change or download. Audit availability depends on the configured audit behavior and retention, so verify the event before promising a particular record to an auditor.
Diagnose by Symptom¶
| Symptom | Safe checks | Resolution and verification |
|---|---|---|
| User sees folder but not document | Check document ACL, inherited profile, active dates, groups, and publication state. | Correct the intended rule; retest with allowed and denied users. |
| Save works but file cannot be downloaded | Check the selected version/file, SQL/archive connectivity, repository availability, service credentials, and logs. | Restore the failing storage dependency; download the same file and an older version. |
| New files work but old files fail | Compare storage locations, archive connectivity, replication status, and encryption configuration. | Restore the missing content/configuration from the matching recovery point; verify both ages. |
| Preview fails but download works | Compare file format and viewer service errors. | Resolve the viewer issue; open the same attachment again. |
| Online editor fails | Check edit permissions, configured WOPI service, discovery, and reachability. | Reopen, edit, save, and read back a test file. |
| Search misses a recent document | Check full-text jobs, service health, extraction failures, and reader ACL. | Resolve the cause; confirm indexing and search as the intended reader. |
| Delete is denied although editing works | Check document Delete separately from Write and document_write. | Change only the intended permission; retest the operation on a disposable version. |
Collect IDs, file names, timestamps, job state, and error messages before escalation. Avoid copying document contents, credentials, or encryption secrets into diagnostic tickets.