Reimplement DME sync as sequential process

• mappings first, datamodels second - datamodel enrichtment (indexing, presentation etc.) depends on mappings
• use single thread executor
• mimic autocrawl schedule behavior