Configuring YARN resources for MapReduce conversion jobs
Cloudera Storage Optimizer uses MapReduce jobs for data conversion. Learn how to configure sufficient YARN resources for MapReduce conversion jobs.
| Daily conversion volume | Concurrent mappers | YARN memory | YARN vCores | Estimated time for conversion |
|---|---|---|---|---|
| Less than 10 TB | 10 | 10 GB | 10 | 2 to 4 hours |
| 10 to 50 TB | 20 | 20 GB | 20 | 3 to 6 hours |
| 50 to 100 TB | 30 | 30 GB | 30 | 4 to 8 hours |
| Greater than 100 TB | 50 | 50 GB | 50 | 6 to 12 hours |
Minimum requirements are as follows:
- Container Memory: 1 GB per mapper
- Container vCores: 1 vCore per mapper
- Concurrent Mappers: Default value is10 (configurable through UI setting Key Conversion Concurrent Mappers)
- Total Minimum: 10 GB RAM and 10 vCores available in YARN
