[HBASE-19021] Restore a few important missing logics for balancer in 2.0

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 2.0.0-alpha-4, 2.0.0
• Component/s: Balancer
• Labels: balancer
• Hadoop Flags: Reviewed

Release Note

Re-enabled 'hbase.master.loadbalance.bytable', default 'false'.
Draining servers are removed from consideration by blancer.balanceCluster() call.

• Assignee: Jerry He
• Reporter: Jerry He
• Created: 16/Oct/17 23:49
• Updated: 22/Apr/19 21:46
• Resolved: 24/Oct/17 17:36

Description

After looking at the code, and some testing, I see the following things are missing for balancer to work properly after AMv2.

1. hbase.master.loadbalance.bytable is not respected. It is always 'bytable'. Previous default is cluster wide, not by table.
2. Servers with no assignments is not added for balance consideration.
3. Crashed server is not removed from the in-memory server map in RegionStates, which affects balance.
4. Draining marker is not respected when balance.

Also try to re-enable TestRegionRebalancing, which has a testRebalanceOnRegionServerNumberChange