-
Notifications
You must be signed in to change notification settings - Fork 1.2k
关于校验数据部分的问题 #21575
Copy link
Copy link
Open
Labels
contributionThis PR is from a community contributor.This PR is from a community contributor.first-time-contributorIndicates that the PR was contributed by an external member and is a first-time contributor.Indicates that the PR was contributed by an external member and is a first-time contributor.
Metadata
Metadata
Assignees
Labels
contributionThis PR is from a community contributor.This PR is from a community contributor.first-time-contributorIndicates that the PR was contributed by an external member and is a first-time contributor.Indicates that the PR was contributed by an external member and is a first-time contributor.
File: /release-8.5/replicate-between-primary-and-secondary-clusters.md
原文档 :
(可选)校验数据。
通过 sync-diff-inspector 工具,可以验证上下游数据在某个时间点的一致性。从上述备份和恢复命令的输出可以看到,上游集群备份的时间点为 431434047157698561,下游集群完成数据恢复的时间点为 431434141450371074。
问题描述:
在多次测试中发现,如果”下游集群完成数据恢复的时间点“,选择restore完成后系统返回那个clusterTs, 大概率是要报”表不存在“的错误。
如果下游集群无其它变更操作,可使用过clusterTs+N秒的Tso,或者就是当前时刻。
[data-sources.downstream]
host = " " # 替换为实际下游集群 ip
port = 4000
user = "root"
password = ""
snapshot = "431434141450371074" # 推荐配置为当前时刻TSO, SELECT (UNIX_TIMESTAMP(now()) * 1000) << 18 AS tso;
原文中的命令:sync_diff_inspector -C ./config.yaml
应改为: tiup sync-diff-inspector -C ./config.yaml