3 ms·
Theoretically that should just be a data sync while maintaining double-write read-primary, and then delete the data from the nodes you don't need anymore once t
by AtlasBarfed 8y ago
Theoretically that should just be a data sync while maintaining double-write read-primary, and then delete the data from the nodes you don't need anymore once the data has been synced? Of course with non-hash indexes the deletions start to slow down with size...
I'm assuming joins, indexes, etc are all isolated to the shard data?