From e9f4b5ab7e21629cc49c8a5cfb1e9ea85ceada6c Mon Sep 17 00:00:00 2001
From: chihe <chihe@oppo.com>
Date: Fri, 14 Jun 2024 18:44:19 +0800
Subject: [PATCH] fix(master): remove the status of DecommissionNeedManualFix

Signed-off-by: chihe <chihe@oppo.com>
---
 master/data_node.go      |  6 +-----
 master/data_partition.go | 11 -----------
 master/disk_manager.go   |  7 +------
 proto/admin_proto.go     |  1 +
 4 files changed, 3 insertions(+), 22 deletions(-)

diff --git a/master/data_node.go b/master/data_node.go
index cdffb8ae6..a04f7394b 100644
--- a/master/data_node.go
+++ b/master/data_node.go
@@ -405,10 +405,6 @@ func (dataNode *DataNode) updateDecommissionStatus(c *Cluster, debug bool) (uint
 			failedNum++
 			failedPartitionIds = append(failedPartitionIds, dp.PartitionID)
 		}
-		if dp.GetDecommissionStatus() == DecommissionNeedManualFix {
-			failedNum++
-			failedPartitionIds = append(failedPartitionIds, dp.PartitionID)
-		}
 		if dp.GetDecommissionStatus() == DecommissionRunning {
 			runningNum++
 			runningPartitionIds = append(runningPartitionIds, dp.PartitionID)
@@ -473,7 +469,7 @@ func (dataNode *DataNode) GetDecommissionFailedDPByTerm(c *Cluster) []proto.Fail
 	partitions := dataNode.GetLatestDecommissionDataPartition(c)
 	log.LogDebugf("action[GetDecommissionDataNodeFailedDP] partitions len %v", len(partitions))
 	for _, dp := range partitions {
-		if dp.IsRollbackFailed() || dp.GetDecommissionStatus() == DecommissionNeedManualFix {
+		if dp.IsRollbackFailed() {
 			failedDps = append(failedDps, proto.FailedDpInfo{PartitionID: dp.PartitionID, ErrMsg: dp.DecommissionErrorMessage})
 			log.LogWarnf("action[GetDecommissionDataNodeFailedDP] dp[%v] failed", dp.PartitionID)
 		}
diff --git a/master/data_partition.go b/master/data_partition.go
index 7c186efa2..b9954ea10 100644
--- a/master/data_partition.go
+++ b/master/data_partition.go
@@ -978,8 +978,6 @@ func GetDecommissionStatusMessage(status uint32) string {
 		return "Success"
 	case DecommissionFail:
 		return "Failed"
-	case DecommissionNeedManualFix:
-		return "DecommissionNeedManualFix"
 	case DecommissionPrepare:
 		return "DecommissionPrepare"
 	default:
@@ -1050,9 +1048,6 @@ func (partition *DataPartition) MarkDecommissionStatus(srcAddr, dstAddr, srcDisk
 	if partition.IsDiscard {
 		goto directly
 	}
-	if partition.needManualFix() {
-		return proto.ErrAllReplicaUnavailable
-	}
 	// TODO-chi:can delete this block
 	// 1 or 2 replica can always add new replica if retrying decommission
 	// for 3 replica,
@@ -1081,7 +1076,6 @@ func (partition *DataPartition) MarkDecommissionStatus(srcAddr, dstAddr, srcDisk
 			if partition.getReplicaDiskErrorNum() == partition.ReplicaNum {
 				log.LogWarnf("action[MarkDecommissionStatus] dp[%v] all replica is unavaliable, cannot handle in auto decommission mode",
 					partition.PartitionID)
-				partition.SetDecommissionStatus(DecommissionNeedManualFix)
 				partition.DecommissionErrorMessage = "all replica is unavailable, cannot handle in auto decommission mode"
 				return proto.ErrAllReplicaUnavailable
 			}
@@ -1121,7 +1115,6 @@ func (partition *DataPartition) MarkDecommissionStatus(srcAddr, dstAddr, srcDisk
 			if partition.getReplicaDiskErrorNum() == partition.ReplicaNum {
 				log.LogWarnf("action[MarkDecommissionStatus] dp[%v] all replica is unavaliable, cannot handle in manual decommission mode",
 					partition.PartitionID)
-				partition.SetDecommissionStatus(DecommissionNeedManualFix)
 				partition.DecommissionErrorMessage = "all replica is unavailable, cannot handle in manual decommission mode"
 				return proto.ErrAllReplicaUnavailable
 			}
@@ -1972,10 +1965,6 @@ func (partition *DataPartition) getDiskErrorReplica() *DataReplica {
 	return nil
 }
 
-func (partition *DataPartition) needManualFix() bool {
-	return partition.GetDecommissionStatus() == DecommissionNeedManualFix
-}
-
 func (partition *DataPartition) checkReplicaMeta(c *Cluster) (err error) {
 	var auditMsg string
 
diff --git a/master/disk_manager.go b/master/disk_manager.go
index 33eebd048..08a08f079 100644
--- a/master/disk_manager.go
+++ b/master/disk_manager.go
@@ -351,11 +351,6 @@ func (dd *DecommissionDisk) updateDecommissionStatus(c *Cluster, debug bool) (ui
 			failedPartitionIds = append(failedPartitionIds, dp.PartitionID)
 		}
 
-		if dp.GetDecommissionStatus() == DecommissionNeedManualFix {
-			failedNum++
-			failedPartitionIds = append(failedPartitionIds, dp.PartitionID)
-		}
-
 		if dp.GetDecommissionStatus() == DecommissionRunning {
 			runningNum++
 			runningPartitionIds = append(runningPartitionIds, dp.PartitionID)
@@ -427,7 +422,7 @@ func (dd *DecommissionDisk) GetDecommissionFailedDPByTerm(c *Cluster) []proto.Fa
 	var failedDps []proto.FailedDpInfo
 	log.LogDebugf("action[GetDecommissionFailedDPByTerm] partitions len %v", len(partitions))
 	for _, dp := range partitions {
-		if dp.IsRollbackFailed() || dp.GetDecommissionStatus() == DecommissionNeedManualFix {
+		if dp.IsRollbackFailed() {
 			failedDps = append(failedDps, proto.FailedDpInfo{PartitionID: dp.PartitionID, ErrMsg: dp.DecommissionErrorMessage})
 			log.LogWarnf("action[GetDecommissionFailedDPByTerm] dp[%v] failed", dp.PartitionID)
 		}
diff --git a/proto/admin_proto.go b/proto/admin_proto.go
index 0a481635b..acd340455 100644
--- a/proto/admin_proto.go
+++ b/proto/admin_proto.go
@@ -47,6 +47,7 @@ const (
 	AdminQueryDataPartitionDecommissionStatus = "/dataPartition/queryDecommissionStatus"
 	AdminCheckReplicaMeta                     = "/dataPartition/checkReplicaMeta"
 	AdminRecoverReplicaMeta                   = "/dataPartition/recoverReplicaMeta"
+	AdminRecoverDiskErrorReplica              = "/dataPartition/recoverDiskErrorReplica"
 	AdminDeleteDataReplica                    = "/dataReplica/delete"
 	AdminAddDataReplica                       = "/dataReplica/add"
 	AdminDeleteVol                            = "/vol/delete"