Wirekite Target Sync
Make the target match the source. In one pass.
Target Sync compares an already-populated target against its source and applies only the differences — the rows that are missing, the rows that drifted, the rows that shouldn't be there. One pass, no CDC, repeatable any time.
When the target drifts.
Targets fall out of sync. A change feed drops for an hour. Someone edits the warehouse by hand. A load stops halfway. Now the target is close to the source — but not exactly equal, and you can't say which rows are wrong.
The usual fixes are both bad. Reload the whole table and you pay for the 99% that was already correct. Stand up a continuous pipeline and you've built a streaming system for a one-time repair.
Target Sync does the third thing: it compares the two databases, finds exactly what differs, and fixes only that. The source is the truth; the target is reconciled to it.
Diff, then apply.
Diff
Compare every row
Wirekite walks the source and the target side by side, table by table, and finds every row that differs — matched by primary key.
Apply
Fix only what's wrong
It inserts the rows that are missing, updates the ones that drifted, and deletes the ones that shouldn't be there — and leaves every row that already matched untouched.
Done
Target equals source
One pass, and the target is an exact copy of the source. Run it again any time the two drift apart.
Which mode do you need?
Data fills an empty target, Replicate keeps it live, Target Sync brings a drifted one back.
| Capability | Wirekite Data | Wirekite Replicate | Target Sync |
|---|---|---|---|
| The job | First load into an empty target | Stream changes continuously | Reconcile a populated target |
| Runs | Once | Always on | On demand, repeatable |
| Target starts | Empty | Already loaded | Loaded, but drifted |
| Uses CDC | No | Yes | No |
| Best for | The initial migration | Keeping a warehouse live | Drift repair and top-ups |
When to reach for it.
Drift repair
A target that diverged from its source — a change feed that dropped, a manual edit, a load that stopped halfway. One pass brings it back into agreement.
Backfill & top-up
Seed or top up a target that has no change feed. Bring it current without standing up a continuous pipeline for a one-time job.
One-shot migration
Move a database where continuous replication isn't available or isn't worth configuring. A single reconciling pass lands the data.
What it needs.
A few honest requirements, up front.
- A primary key on every table
- Updates and deletes are matched by primary key, so each table in scope needs one.
- The target schema already exists
- Target Sync moves rows, not table definitions. If the target is empty, migrate the schema first, then sync the data into it.
- One direction per run
- The source is the source of truth and the target is reconciled to it. To sync the other way, swap the two and run it again.
- A supported source
- Postgres, MySQL, MariaDB, SQL Server, Oracle, and Cassandra can drive a Target Sync. The target can be any database or warehouse Wirekite writes to.
- Aligned key case-sensitivity
- The source and target should agree on whether primary keys are case-sensitive, so the row-by-row match lines up cleanly.
Every supported pair
See which engines support Target Sync.
The full matrix shows every engine as a source, as a target, and in each of the three modes — Data, Replicate, and Target Sync.
Open the matrix
Have a target that's drifted?
We'll run a reconciling pass on a sample of your workload during the demo — and show you exactly what changed.