diff --git a/datamover/pkg/drivers/https/migration.go b/datamover/pkg/drivers/https/migration.go
index 200192a5d..4fa1745a1 100644
--- a/datamover/pkg/drivers/https/migration.go
+++ b/datamover/pkg/drivers/https/migration.go
@@ -27,7 +27,7 @@ import (
 
 var simuRoutines = 10
 var PART_SIZE int64 = 16 * 1024 * 1024 //The max object size that can be moved directly, default is 16M.
-var JOB_RUN_TIME_MAX = 86400           //seconds
+var JOB_RUN_TIME_MAX = 86400           //seconds, equals 1 day
 var s3client osdss3.S3Service
 var bkendclient backend.BackendService
 
@@ -360,6 +360,10 @@ func multipartMoveObj(obj *osdss3.Object, srcLoca *LocationInfo, destLoca *Locat
 	err = completeMultipartUpload(uploadObjKey, destLoca, uploadMover)
 	if err != nil {
 		logger.Println(err.Error())
+		err := abortMultipartUpload(obj.ObjectKey, destLoca, uploadMover)
+		if err != nil {
+			logger.Printf("Abort s3 multipart upload failed, err:%v\n", err)
+		}
 	}
 
 	return err
@@ -679,7 +683,19 @@ func runjob(in *pb.RunJobRequest) error {
 	ctx := context.Background()
 	_, ok := ctx.Deadline()
 	if !ok {
-		ctx, _ = context.WithTimeout(ctx, 7200*time.Second)
+		tmoutCfg, err := strconv.ParseInt(os.Getenv("JOB_MAX_RUN_TIME"), 10, 64)
+		if err != nil || tmoutCfg < 60 || tmoutCfg > 2592000 {
+			tmoutCfg = int64(JOB_RUN_TIME_MAX)
+		}
+		durStr := fmt.Sprintf("%ds", tmoutCfg)
+		logger.Printf("Vaule of JOB_MAX_RUN_TIME is: %d seconds, durStr:%s.\n", tmoutCfg, durStr)
+		tmout, err := time.ParseDuration(durStr)
+		if err == nil {
+			ctx, _ = context.WithTimeout(ctx, tmout)
+		} else {
+			logger.Println("Set timeout to the default value.")
+			ctx, _ = context.WithTimeout(ctx, 86400*time.Second) //1 day as default
+		}
 	}
 
 	//prepare for running
diff --git a/datamover/pkg/hw/obs/obsmover.go b/datamover/pkg/hw/obs/obsmover.go
index 7a74516b3..763809e08 100644
--- a/datamover/pkg/hw/obs/obsmover.go
+++ b/datamover/pkg/hw/obs/obsmover.go
@@ -232,6 +232,7 @@ func (mover *ObsMover) UploadPart(objKey string, destLoca *LocationInfo, upBytes
 	uploadPartInput.Offset = offset
 	uploadPartInput.PartSize = upBytes
 	for tries := 1; tries <= 3; tries++ {
+		uploadPartInput.Body = bytes.NewReader(buf)
 		uploadPartInputOutput, err := mover.obsClient.UploadPart(uploadPartInput)
 		if err != nil {
 			log.Logf("[obsmover] Upload object[%s] range[partnumber#%d,offset#%d] failed %d times. err:%v\n",
diff --git a/docker-compose.yml b/docker-compose.yml
index a3e75c3d2..60840d368 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -66,7 +66,10 @@ services:
       - MICRO_REGISTRY=mdns
       - DB_HOST=datastore:27017
       - KAFKA_ADVERTISED_LISTENERS=PLAINTEXT://kafka:9092
+      #PARTSIZE for multi-part migration, should not be less than 5(MB) or more than 100(MB)
       - PARTSIZE=16
+      #JOB_MAX_RUN_TIME for each migration job, should not be less than 60(second) or more than 2592000(second)
+      - JOB_MAX_RUN_TIME=86400
     depends_on:
       - zookeeper
       - kafka