--- title: "OceanBase 数据库 V4.3 版本中集群合并出现 tablet haven't kept medium snapshot(ret=-4016 的原因及处理方法-OceanBase数据库使用指南" description: "了解OceanBase数据库在实际应用中关于 OceanBase 数据库 V4.3 版本中集群合并出现 tablet haven't kept medium snapshot(ret=-4016 的原因及处理方法相关的常见问题和使用技巧，帮助您快速解决 OceanBase 数据库 V4.3 版本中集群合并出现 tablet haven't kept medium snapshot(ret=-4016 的原因及处理方法的难题。" --- 切换语言 - 中文站 - 简体中文 - International - English - 日本站 - 日本語 划线反馈 # OceanBase 数据库 V4.3 版本中集群合并出现 tablet haven't kept medium snapshot(ret=-4016 的原因及处理方法 更新时间：2025-03-02 06:11 适用版本： V4.0.x、V4.1.x、V4.2.x、V4.3.x 内容类型：Troubleshoot ## 问题现象 OceanBase 数据库 V4.3 版本中集群合并卡住，部分 tablet 合并出现 -4016 异常。 ```shell obclient> select * from oceanbase.__all_virtual_compaction_diagnose_info; ``` 输出关键结果如下： ![image001](https://obbusiness-private.oss-cn-shanghai.aliyuncs.com/doc/img/knowledge-base/database/transfer-compaction/compaction/20241106ob433compaction001.png) 日志伴随如下异常信息。 ```shell observer.log.20241024095020440:[2024-10-24 09:48:21.961621] WDIAG [STORAGE] get_medium_merge_tables (ob_partition_merge_policy.cpp:106) [1453611][T1002_CO_MERGE_][T1002][YBxxxxx-xxxxx-xxxxx-xxxxx] [lt=77][errcode=-4016] tablet haven't kept medium snapshot(ret=-4016, tablet={this:0x7fe1ebd4f7e0, is_inited:true, wash_score:1729734501951129000, hold_ref_cnt:true, ref_cnt:2, version:4, length:0, tablet_addr:{[ver=1,mode=0,seq=1232291][2nd=1119785][3rd=0][offset=643072,size=773012,type=4,seq=0][5th=0][6th=0]}, allocator:null, tablet_meta:{version:1, ls_id:{id:1002}, tablet_id:{id:205660}, data_tablet_id:{id:205660}, ref_tablet_id:{id:0}, has_next_tablet:false, create_scn:{val:18446744073709551615, v:3}, start_scn:{val:1, v:0}, clog_checkpoint_scn:{val:1729301277271099048, v:0}, ddl_checkpoint_scn:{val:1, v:0}, snapshot_version:1729301276444306053, multi_version_start:1729272306299137000, compat_mode:1, ha_status:{restore_status:0, data_status:0, expected_status:0, reserved:0}, report_status:{merge_snapshot_version:1729188003129171000, cur_report_version:0, data_checksum:0, row_count:0, found_cg_checksum_error:false}, table_store_flag:{with_major_sstable:1}, ddl_start_scn:{val:0, v:0}, ddl_snapshot_version:0, max_sync_storage_schema_version:0, max_serialized_medium_scn:0, ddl_execution_id:-1, ddl_data_format_version:0, ddl_commit_scn:{val:0, v:0}, mds_checkpoint_scn:{val:1729301579520857021, v:0}, transfer_info:{ls_id:{id:0}, transfer_start_scn:{val:0, v:0}, transfer_seq:0, has_transfer_table:false}, extra_medium_info:{info:4097, compat:1, last_compaction_type:0, wait_check_flag:1, last_medium_scn:1729188003129171000}, last_persisted_committed_tablet_status:{tablet_status:{val:1, str:"NORMAL"}, transfer_scn:{val:18446744073709551615, v:3}, transfer_ls_id:{id:-1}, data_type:1, create_commit_scn:{val:1729257834170270107, v:0}, create_commit_version:1729257834049127004, delete_commit_scn:{val:18446744073709551615, v:3}, delete_commit_version:-1, transfer_out_commit_version:-1}, create_schema_version:1729257830809200, space_usage:{all_sstable_data_occupy_size:49118775722, all_sstable_data_required_size:50608472064, all_sstable_meta_size:10485760, ss_public_sstable_occupy_size:0, tablet_clustered_meta_size:794624, tablet_clustered_sstable_data_size:0, backup_bytes:0}, micro_index_clustered:false, ddl_table_type:14}, table_store_addr:{ptr:0x7fe1ebd4fe04, addr:{[ver=1,mode=0,seq=1232260][2nd=1119754][3rd=0][offset=1220608,size=613,type=2,seq=0][5th=0][6th=0]}}, storage_schema_addr:{ptr:null, addr:{[ver=1,mode=0,seq=1232260][2nd=1119754][3rd=0][offset=1221221,size=1067,type=2,seq=0][5th=0][6th=0]}}, macro_info_addr:{ptr:null, addr:{[ver=1,mode=0,seq=1232291][2nd=1119785][3rd=0][offset=643572,size=772512,type=4,seq=0][5th=0][6th=0]}}, mds_data:null, ddl_kvs:null, ddl_kv_count:0, is_external_tablet:false, table_store_cache:{last_major_snapshot_version:1729188003129171000, major_table_cnt:1, minor_table_cnt:2, recycle_version:1729188003129171000, last_major_column_count:31, is_row_store:false, last_major_compressor_type:1, last_major_latest_row_store_type:1}, rowkey_read_info:0x7fe1ebd4fc00, is_cs_replica_compat:false, memtables:[null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null], memtable_count:0}, param={merge_type:"MEDIUM_MERGE", merge_version:1729258011030261005}) ``` ## 问题原因 OceanBase 数据库内部在处理 clog 回放时逻辑存在异常导致。 ## 问题的风险及影响 集群合并异常。 ## 影响租户 影响 OceanBase 数据库中的 SYS 租户和 Oracle 租户以及 MySQL 租户。 ## 适用版本 OceanBase 数据库 V4.x 版本。 ## 解决方法 - 彻底解决方法： 升级至问题已修复版本。目前已修复的版本包括 OceanBase 数据库 V4.3.3 BP1（oceanbase-4.3.3.1-101000032024102022）及之后版本。 - 问题出现后临时解决方法： 通过如下命令将异常的日志流删除，删除后 OceanBase 数据库会自动补足日志流，后合并会自动恢复，示例如下。 ```shell ALTER SYSTEM REMOVE REPLICA LS = 1001 SERVER = 'xxx.xxx.xxx.001:2882' TENANT = 'mysql001'; ``` 上一篇 [合并期间空间膨胀问题](https://www.oceanbase.com/knowledge-base/oceanbase-database-1000000000209907) 下一篇 [enable_manual_merge 参数是否会影响 alter system major freeze 的执行结果](https://www.oceanbase.com/knowledge-base/oceanbase-database-1000000000267166) ![有帮助](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*y6ocSqN8cqsAAAAAAAAAAAAAARQnAQ)![无帮助](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*BG9IQJyLHF8AAAAAAAAAAAAAARQnAQ)![反馈](https://gw.alipayobjects.com/mdn/ob_asset/afts/img/A*eTWdQKCRKHwAAAAAAAAAAAAAARQnAQ)[AI](https://www.oceanbase.com/obi) 咨询热线