创建索引报错

【详述】创建布隆索引报错
【背景】ALTER TABLE yl_jms_spmi_transfer_bill_pt SET (“bloom_filter_columns” = “task_code”);
【业务影响】
【StarRocks版本】例如:2.5.10
【集群规模】例如:3fe(1 follower+2observer)+3be
【机器信息】CPU虚拟核/内存/网卡,例如:48C/64G/万兆

【附件】

schema change task failed: task type: ALTER, status_code: RUNTIME_ERROR, backendId: Backend [id=10003, host=10.33.81.49, heartbeatPort=9050, alive=true], signature: 57529

show tablet 53453
-1 -1 -1 -1 false SHOW PROC ‘/dbs/-1/-1/partitions/-1/-1/53453’;

be.WARNING.log.20230821-162758.lz4 (75.6 MB)

这个表的数据量有多大,一个节点几块盘,请发下leader 节点 fe.log

fe.log.20230821-6.lz4 (38.3 MB) 190亿,12T数据,6块盘。fe.log.20230821-4.lz4 (12.5 MB) fe.log.20230821-5.lz4 (12.6 MB)

可以调大be.conf 中的 alter_tablet_worker_count ,默认值是3,可以调大观察下

请问下,加索引的过程中,会查询不了吗?
1064 - Build Exec OlapScanNode fail, scan info is invalid,Failed to get scan range, no queryable replica found in tablet: 40056 10005:27332/-1/27332/27331,10004:27332/-1/27332/27331,10003:27332/-1/27332/27331,

调整到18,跑了差不多3个小时,还是报错了
W0822 15:08:07.113236 4239 agent_task.cpp:122] alter failed. signature: 76939
W0822 15:08:07.113606 4239 schema_change.cpp:721] fail to find new tablet. new_tablet=72847, new_schema_hash=276542630
W0822 15:08:07.113619 4239 engine_alter_tablet_task.cpp:57] failed to do alter task. status=Internal error: failed to find new tablet base_tablet_id=38168, base_schema_hash=1170333633, new_tablet_id=72847, new_schema_hash=276542630

麻烦发一下be的info日志,或者您这边有联系方式吗?我是SR这边的研发,我加下您了解一些信息。

您的FE建议换成3台follower,实现高可用

这个问题有解决吗?我们也遇到了这个