Recovery
If a conversion is interrupted (crash, power loss, etc.), Pebblify can resume from the last checkpoint.
Resume a Conversion
pebblify recover [options]
Examples
pebblify recover
pebblify recover --tmp-dir /var/tmp
Options
| Flag | Description |
|---|---|
-w, --workers N | Max concurrent DB conversions (0 = auto) |
-v, --verbose | Enable verbose output |
--tmp-dir DIR | Directory containing .pebblify-tmp/ |
--health | Enable HTTP health probe server |
--health-port P | Health probe port (default: 8086) |
--metrics | Enable Prometheus metrics exporter |
--metrics-port P | Prometheus metrics port (default: 9090) |
How Recovery Works
Pebblify maintains a state file at .pebblify-tmp/state.json that tracks:
- Source and destination paths
- Status of each database (pending, in_progress, done, failed)
- Last checkpoint key for each database
- Migration statistics and metrics
When you run recover, Pebblify reads this state file and resumes conversion from the last successfully checkpointed key in each database, skipping any databases that were already completed.
tip
Make sure to use the same --tmp-dir value as the original conversion, otherwise Pebblify will not find the state file.