8 hour rebalance with a swap rebalance of a single node


#1

I’ve never seen this behaviour, rebalances in our 2 group, 1 replica environment takes 45 minutes to move 1/12 of 800M objects. However I’m seeing this same process (swap rebalance take 8 hours in a new environment that has 3 groups and 2 replicas). Trying to figure out if the new 3 group, 2 replica is causing the outlandish times or what. No errors and it does complete…

cluster 4.6.4 (don’t ask, we can’t move to 5 yet)
CentOS
12 nodes
800million objects.

Again normally this takes 45 minutes in the same environment, but with 2 groups and only 1 replica.