Closed mboysan closed 5 years ago
If performance of the overall database will be low because of this, I will fix it. But due to the complexity of the task, need to profile the db first to see the bottleneck before applying this fix.
It turned out to be a critical part of the Bizur algorithm since the synchronization of operations are done per bucket and would cause serious blocking problems. Therefore, it is fixed.
Leaders should be assigned per bucket, not per node.