修复副本失败,clone失败

【详述】修复副本一直失败,提示clone failed
【背景】测试be挂掉一个以后,继续写数据成功,ADMIN SHOW REPLICA STATUS FROM tablename; 查看有副本不健康,然后使用ADMIN SET REPLICA STATUS PROPERTIES(“tablet_id” = “15147”, “backend_id” = “10008”, “status” = “bad”); 设置为bad,过会开始修复 clone数据,但是反反复复一直没成功,查看be.WARNING报错
【业务影响】
【StarRocks版本】2.5.4
【集群规模】1fe + 3be
【机器信息】CPU虚拟核/内存/网卡,例如:8C/16G/万兆
【联系方式】社区群13-妈妈说名字起太长点赞时候能占一行,谢谢
【附件】

  • be.warning/相应截图
    W0530 15:41:54.903019 9632 engine_clone_task.cpp:161] Fail to load snapshot:Runtime error: Fail to link segment data file tablet:10089
    W0530 15:41:54.903059 9632 agent_task.cpp:321] clone failed. signature: 10089
    W0530 15:41:58.893393 9632 tablet_manager.cpp:960] Fail to remove or move /opt/StarRocks-2.5.4/be/storage/data/840/15147 :IO error: /opt/StarRocks-2.5.4/be/storage/data/840/15147: Invalid cross-device link
    W0530 15:41:58.893429 9632 tablet_manager.cpp:998] before adding new cloned tablet, delete stale TABLET_SHUTDOWN tablet failed after 0 times retry, tablet:15147 st:IO error: /opt/StarRocks-2.5.4/be/storage/data/840/15147: Invalid cross-device link
    W0530 15:41:58.893445 9632 agent_task.cpp:321] clone failed. signature: 15147

你好 是几副本的表 麻烦show tablet $tabletid看下

show tablet 15147;


SHOW PROC ‘/dbs/10157/15145/partitions/15144/15146/15147’; 一般是出现三个,有一个在clone,这会有变为两个了

ADMIN SHOW REPLICA STATUS FROM mem_individual_car; 现在这个sql不能用了,估计待会可以查

ADMIN SHOW REPLICA STATUS FROM TABLENAME;

解决了,运维挂了块盘在data目录,和其他目录不在一个卷,现在盘移动到/opt/StarRocks-2.5.4 目录就好了,副本修复了