跨实例备份恢复milvus时报错

[IeohMingChan]

根据官网指示https://milvus.io/docs/zh/cross-bucket-backup-and-restore.md 进行跨存储桶实例迁移，在恢复阶段报错：restore collection failed，已进行多次尝试，包括指定collection，db，后缀等，具体情况如下：
版本号：v2.4.1 集群版
待备份数据结构如下，备份图中的所有集合：

备份过程：

1. 修改配置文件关键信息如下：

milvus:
address: xxx
port: 30000
user: "xxx"
password: "xxx"

minio:
storageType: "minio" # support storage type: local, minio, s3, aws, gcp, ali(aliyun), azure, tc(tencent)
address: xxx # Address of MinIO/S3
port: 30001 # Port of MinIO/S3
accessKeyID: xxx # accessKeyID of MinIO/S3
secretAccessKey: xxx # MinIO/S3 encryption string
useSSL: false # Access to MinIO/S3 with SSL
useIAM: false
iamEndpoint: ""
bucketName: "paas-milvus" # Milvus Bucket name in MinIO/S3, make it the same as your milvus instance
rootPath: "" # Milvus storage root path in MinIO/S3, make it the same as your milvus instance

backupStorageType: "minio" # support storage type: local, minio, s3, aws, gcp, ali(aliyun), azure, tc(tencent)
backupAddress: xxx # Address of MinIO/S3
backupPort: 30001 # Port of MinIO/S3
backupAccessKeyID: xxx # accessKeyID of MinIO/S3
backupSecretAccessKey:xxx # MinIO/S3 encryption string
backupBucketName: "test1-milvus" # Bucket name to store backup data. Backup data will store to backupBucketName/backupRootPath
backupRootPath: "backup" # Rootpath to store backup data. Backup data will store to backupBucketName/backupRootPath
crossStorage: "false"

2. 检查配置：./milvus-backup check。结果正常
3. 执行备份操作备份全部数据（也尝试过备份指定集合或数据库）：./milvus-backup create -n my_backup。结果正常
4. 修改配置文件准备恢复：
   milvus.port改为30010，即指向新的milvus地址
   minio.bucketName改为"test1-milvus"
5. 执行恢复操作：./milvus-backup restore -n my_backup。期间发生报错，主要内容为restore coll failed，restore collection failed，execute restore collection fail和bulk insert failed: Wrong rows type: invalid parameter[expected=SparseFloatVectorFieldData]。具体情况见下图：
   
   
6. 执行结束后，目标milvus中只有两个集合且无数据，加载按钮不可点击
   

请教下如何修复该问题，万分感谢！

[yhmo]

sparse vector是milvus 2.4.0开始有的功能。备份恢复要调用bulk_insert这个接口导入数据文件，而bulkinsert接口支持sparse vector是在milvus 2.4.3才有：

所以你的milvus 2.4.1会在导入含有sparse vector的数据文件时报错，然后整个恢复任务就失败。
建议尽量使用每个大版本的后期小版本。

[IeohMingChan]

感谢指导，现从2.4.1版本备份数据并恢复至2.5.5版本中，暂时表现正常