← TATER Help · Endpoint troubleshooting

OneDrive Business1 regression diagnostic

Some users hit a recurring failure where the OneDrive Business1 registry record is present but the UserFolder value is blank. Files won't hydrate because cldflt has no upstream binding. The most painful variant: the regression returns within 7 days of a full uninstall+reinstall repair, with no obvious user action between fix and breakage. Tracked as ADO #583. First captured on a Caron Bletzer host (2026-06-03).

Signature

Automated detection

As of agent v2.4.21+, the OneDrive health monitor reports this as a distinct signal: od.business1.userfolder-blank (fail grade). It is mapped to remediation REM_OD_011_DiagnoseRegression rather than a destructive repair — the field evidence says no current repair survives, so the first action is to capture forensic state instead of compounding the problem.

The signal raises a MonitoringFinding with source=onedrive, resource=Business1 on <hostname>. Trigger the diagnostic via the standard MonitoringFindings promote-to-task path; admins get a task with the detailed diagnostic script as the next-step instruction.

Diagnostic capture

Script: Runbooks/Remediations/OneDrive/REM_OD_011_DiagnoseRegression.ps1. Run it as the affected user (preferred) OR via Task Scheduler in LoadUserProfile mode (so HKCU points at the right hive). Outputs a structured JSON report covering:

Each run writes a snapshot to %ProgramData%\TATER\onedrive-snapshots\<user>-<ts>.json (last 30 retained). The agent caller can diff against the previous snapshot to identify the field that mutated between runs — typically the strongest signal for "what is rewriting Business1 in the background."

Exit codes

CodeMeaning
0Healthy — Business1 present and UserFolder populated.
2User profile not loaded (run as the user OR with LoadUserProfile).
3Regression detected — Business1 present, UserFolder blank.

Investigation checklist

When the regression is captured, work through these in order against the JSON output:

  1. Business1 last-write timestamp vs. Day-0 repair timestamp. Confirms the record really was rewritten (vs a UserFolder that never populated after a partial repair).
  2. Group Policy History — any OneDrive-touching GPO that applied in the last 8 days. The CSE GUID is fixed; look for DisplayName / GPOLink fields.
  3. Intune CSP state — compare OneDriveNGSC values across snapshots. The "Silent account configuration" CSP has historically reset UserFolder when its tenant guidance changes.
  4. Scheduled tasks — OneDrive update / Standalone Update Task / Sign-in helper. Cross-reference LastRunUtc with the Business1 last-write timestamp; the closest preceding task run is the likely culprit.
  5. RecentEvents — OneDrive event IDs 1, 1000, 7036 (cldflt service start/stop), 1234 (account reset) frame the timeline.
  6. ClientPolicy.ini files — sometimes the ClientPolicy.ini bound to the missing library is deleted, which explains the "no library" symptom.

What the diagnostic does NOT do

Once the cause is known

Add the identified mutation source (GPO name / Intune CSP path / scheduled-task ARN) to the BUSINESS1_REGRESSION_SOURCES list in Agent/go/internal/onedrive/grader.go and ship a Tier 3 remediation that targets it explicitly. The combined-variant signal will auto-clear on the next agent cycle.