Unify the re-balancing logic for `ec.encode` with `ec.balance`. (#6339)

Among others, this enables recent changes related to topology aware
re-balancing at EC encoding time.

1 files changed, 1 insertions, 67 deletions

diff --git a/weed/shell/command_ec_balance.go b/weed/shell/command_ec_balance.go
index 81d214fcc..9e27f40ce 100644
--- a/weed/shell/command_ec_balance.go
+++ b/weed/shell/command_ec_balance.go
@@ -24,73 +24,7 @@ func (c *commandEcBalance) Help() string {
 	ec.balance [-c EACH_COLLECTION|<collection_name>] [-force] [-dataCenter <data_center>] [-shardReplicaPlacement <replica_placement>]
 
 	Algorithm:
-
-	func EcBalance() {
-		for each collection:
-			balanceEcVolumes(collectionName)
-		for each rack:
-			balanceEcRack(rack)
-	}
-
-	func balanceEcVolumes(collectionName){
-		for each volume:
-			doDeduplicateEcShards(volumeId)
-
-		tracks rack~shardCount mapping
-		for each volume:
-			doBalanceEcShardsAcrossRacks(volumeId)
-
-		for each volume:
-			doBalanceEcShardsWithinRacks(volumeId)
-	}
-
-	// spread ec shards into more racks
-	func doBalanceEcShardsAcrossRacks(volumeId){
-		tracks rack~volumeIdShardCount mapping
-		averageShardsPerEcRack = totalShardNumber / numRacks  // totalShardNumber is 14 for now, later could varies for each dc
-		ecShardsToMove = select overflown ec shards from racks with ec shard counts > averageShardsPerEcRack
-		for each ecShardsToMove {
-			destRack = pickOneRack(rack~shardCount, rack~volumeIdShardCount, ecShardReplicaPlacement)
-			destVolumeServers = volume servers on the destRack
-			pickOneEcNodeAndMoveOneShard(destVolumeServers)
-		}
-	}
-
-	func doBalanceEcShardsWithinRacks(volumeId){
-		racks = collect all racks that the volume id is on
-		for rack, shards := range racks
-			doBalanceEcShardsWithinOneRack(volumeId, shards, rack)
-	}
-
-	// move ec shards 
-	func doBalanceEcShardsWithinOneRack(volumeId, shards, rackId){
-		tracks volumeServer~volumeIdShardCount mapping
-		averageShardCount = len(shards) / numVolumeServers
-		volumeServersOverAverage = volume servers with volumeId's ec shard counts > averageShardsPerEcRack
-		ecShardsToMove = select overflown ec shards from volumeServersOverAverage
-		for each ecShardsToMove {
-			destVolumeServer = pickOneVolumeServer(volumeServer~shardCount, volumeServer~volumeIdShardCount, ecShardReplicaPlacement)
-			pickOneEcNodeAndMoveOneShard(destVolumeServers)
-		}
-	}
-
-	// move ec shards while keeping shard distribution for the same volume unchanged or more even
-	func balanceEcRack(rack){
-		averageShardCount = total shards / numVolumeServers
-		for hasMovedOneEcShard {
-			sort all volume servers ordered by the number of local ec shards
-			pick the volume server A with the lowest number of ec shards x
-			pick the volume server B with the highest number of ec shards y
-			if y > averageShardCount and x +1 <= averageShardCount {
-				if B has a ec shard with volume id v that A does not have {
-					move one ec shard v from B to A
-					hasMovedOneEcShard = true
-				}
-			}
-		}
-	}
-
-`
+	` + ecBalanceAlgorithmDescription
 }
 
 func (c *commandEcBalance) HasTag(CommandTag) bool {