Skip to main content

Roll Back a Sync

A sync brought in data you didn't want — a mis-mapped column, a test file, a spreadsheet with the wrong tab selected. Rolling back removes the data that run added, so you can fix the problem at the source and sync again.

What you'll end up with

The records that run created removed from Oho, and the data source left connected and ready to re-run once you've fixed the cause.

Before you start​

Rollback is only offered on the most recent successful run of a data source. If the bad sync has since been followed by another successful one, the option won't appear — the newer run is the only one you can roll back.

Work out what actually went wrong before rolling back. A rollback undoes the import; it doesn't stop the same thing happening on the next scheduled sync. If the cause is a field mapping, fix that first — see Map your fields.

Step 1: Open the data source​

In the left menu under Admin, open Data Sources and select the source whose sync you want to undo.

Step 2: Find the run​

Open the source's run history. Each row is one sync, with when it ran, how long it took, and how many records it touched.

Find the run you want to undo. Check the timestamp against when the bad data appeared — the run list is the quickest way to confirm you're undoing the right one.

Step 3: Roll it back​

Choose Rollback on that run. Oho asks you to confirm:

Rolling back this ingestion run will remove any new data ingested during the run. This may exclude data that was previously extracted, but did not change during this run.

That caveat is the important part. Read the next section before confirming.

What a rollback does and doesn't remove​

A rollback removes what that run added. It does not reset your workforce to how it looked beforehand.

  • Removed — records the run created.
  • Not removed — records that already existed and didn't change during the run. They were not part of what the run wrote, so there is nothing to undo.
  • Not reverted — anything a person edited in Oho after the run.

So if a bad sync updated 40 existing workers and created 5 new ones, rolling it back removes the 5 it created. The 40 keep whatever the run wrote to them. Where a run has overwritten good data with bad, the fix is to correct the source and sync again, not to roll back.

Rollback deletes records

Records the run created are removed, not archived. If you're unsure whether a run is the right one, check the timestamp and record counts in the run history first.

Step 4: Fix the cause, then re-sync​

Rollback clears the bad data; it doesn't prevent a repeat. Before the next scheduled sync:

  1. Fix the problem at its origin — the field mapping, the source file, or the credentials the connector uses.
  2. Run the sync again and check the result, as in Run and verify your first sync.

Common questions​

The Rollback option isn't there. It only appears on a source's most recent successful run. A later successful sync will have taken its place.

Can I roll back several runs? No — one run at a time, and only the latest successful one.

Will it remove workers people have edited since? Records the run created are removed even if they were edited afterwards. Edits to records that already existed are unaffected.