Doris BE 宕机后数据无法导入和查询

Viewed 10

现状
Doris 2.1.9版本。
当前副本数量是单副本

问题
在Doris BE 突然宕机后。 BE 再次启动后,Doris某些表无法正常Doris数据, 也无法正常查询

导入时报错:

[ANALYSIS_ERROR]TStatus: errCode = 2, detailMessage = Failed to commit txn 459678240, cause tablet 740453653 succ replica num 0 < load required replica num 1. table 73862495, partition: [ id=73862494, commit version 625473, visible version 625460 ], this tablet detail: 1 replicas write data succ but miss previous version: { [replicaId=740453654, backendId=10236, backendAlive=true, version=625460, lastFailedVersion=625461, lastSuccessVersion=625460, lastFailedTimestamp=0, state=NORMAL] }, please check output of SHOW TABLET for details. Please try again later.

查询时报错:
ERROR 1105 (HY000): errCode = 2, detailMessage = Failed to get scan range, no queryable replica found in tablet: 520050185. Reason: Visible Replicas:Visible version: 386708, Replicas: [replicaId=520050186, backendId=10236, backendAlive=true, version=386708, lastFailedVersion=386709, lastSuccessVersion=386708, lastFailedTimestamp=0, state=NORMAL].

期望
1、期望Doris 可以修复此bug
2、有什么办法可以紧急的修复数据。(当前是truncate 表重新导入的,数据量小还好,一旦数据量大,就影响很大)

问题同: https://github.com/apache/doris/issues/55754

1 Answers

我也遇到这个问题,是doris-fe报错吗?我的doris-fe报错是

SELECT * FROM `__internal_schema`.`column_statistics` WHERE `id` = '20227--1-product_output' AND `catalog_id` = '0' AND `db_id` = '20173'
org.apache.doris.common.UserException: errCode = 2, detailMessage = Failed to get scan range, no queryable replica found in tablet: 11079. Reason: Visible Replicas:Visible version: 235363, Replicas: [replicaId=11080, backendId=11030, backendAlive=true, version=235363, isBad=true, state=NORMAL].

我手动删除了__internal_schema.column_statistics然后重启了doris-fe
在此之前我使用meta-tool删除了对应的tablet,也到数据目录找到该tablet的位置,手动删除
但是我现在遇到另外一个问题就是doris-be经常重启