starrocks3.3.0表索引文件缺失导致写入失败

为了更快的定位您的问题,请提供以下信息,谢谢
【详述】:晚上海豚调度跑自动任务时插不进去数据
【是否存算分离】:没有存算分离
【StarRocks版本】例如:3.3.0
【集群规模】例如:3be+6fe(fe与be混部)
【机器信息】CPU虚拟核/内存/网卡,例如:32c/128G/万兆

  • fe.log/beINFO/相应截图
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:W0325 01:55:21.441236 30043 tablet_sink_sender.cpp:168] NodeChannel[194742251], tablet open failed, load_id=a0658f70-27aa-11f1-a405-00163e0c9e00, txn_id: 848086831, parallel=1, compress_type=2, node=192.168.0.196:8060, errmsg=Tablet is in error state, tablet_id: 214825198 primary index commit failed: Not found: /data/star_rocks/storage/data/872/214825198/1133836910/index.l2.7631.2.merged: No such file or directory
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:be/src/storage/persistent_index.cpp:3528 _reload(*index_meta) tablet:214825198 #version:38 [7619.1 7634@36 7634.1] pending: rowsets:2[id/seg/row/del/byte/compaction]: [17654/1/84495/0/924.45 KB/255.10 MB],[17655/1/6778173/0/58.08 MB/197.92 MB]
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:W0325 01:55:21.441249 30043 tablet_sink_sender.cpp:177] Open channel failed. load_id: a0658f7027aa11f1-a40500163e0c9e00, error: Service unavailable: Tablet is in error state, tablet_id: 214825198 primary index commit failed: Not found: /data/star_rocks/storage/data/872/214825198/1133836910/index.l2.7631.2.merged: No such file or directory
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:be/src/storage/persistent_index.cpp:3528 _reload(*index_meta) tablet:214825198 #version:38 [7619.1 7634@36 7634.1] pending: rowsets:2[id/seg/row/del/byte/compaction]: [17654/1/84495/0/924.45 KB/255.10 MB],[17655/1/6778173/0/58.08 MB/197.92 MB]: be:192.168.0.196
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:W0325 01:55:36.597323 30046 tablet_sink_sender.cpp:168] NodeChannel[194742251], tablet open failed, load_id=a9713c91-27aa-11f1-a405-00163e0c9e00, txn_id: 848086982, parallel=1, compress_type=2, node=192.168.0.196:8060, errmsg=Tablet is in error state, tablet_id: 214825198 primary index commit failed: Not found: /data/star_rocks/storage/data/872/214825198/1133836910/index.l2.7631.2.merged: No such file or directory
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:be/src/storage/persistent_index.cpp:3528 _reload(*index_meta) tablet:214825198 #version:38 [7619.1 7634@36 7634.1] pending: rowsets:2[id/seg/row/del/byte/compaction]: [17654/1/84495/0/924.45 KB/255.10 MB],[17655/1/6778173/0/58.08 MB/197.92 MB]
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:W0325 01:55:36.597361 30046 tablet_sink_sender.cpp:177] Open channel failed. load_id: a9713c9127aa11f1-a40500163e0c9e00, error: Service unavailable: Tablet is in error state, tablet_id: 214825198 primary index commit failed: Not found: /data/star_rocks/storage/data/872/214825198/1133836910/index.l2.7631.2.merged: No such file or directory
    /data/star_rocks/be/log/be.WARNING.log.20260315-213354:be/src/storage/persistent_index.cpp:3528 _reload(*index_meta) tablet:214825198 #version:38 [7619.1 7634@36 7634.1] pending: rowsets:2[id/seg/row/del/byte/compaction]: [17654/1/84495/0/924.45 KB/255.10 MB],[17655/1/6778173/0/58.08 MB/197.92 MB]: be:192.168.0.196
    阿里云服务器,单副本,内存没超,进程没重启,没人操作删除东西

单副本?

是的,单副本
3BE,6FE 混合部署
服务没重启,磁盘io无异常 内存没异常

单副本确实悬,在不考虑外部因素(如磁盘损坏、人为误删)等的情况下,sr内部也是有可能会导致tablet损坏的,特别是频繁读写高强度的Compaction场景下,只有用多副本,出现这种它才会自愈。否则只能重新导入数据了。