Zookeeper 如何分析ReplicasMaxAbsoluteDelay为什么返回当前时间问题？

zf9nrax1 于 2022-12-09 发布在 Apache

关注(0)|答案(1)|浏览(116)

Sometimes when the zabbix monitor executes the sql "select value from system.asynchronous_metrics where metric='ReplicasMaxAbsoluteDelay';", it will return a very large value like "52y 9m 18d" (currenttime - 1970).
According to the getAbsoluteDelay() function, we knew when queue_update_finish_time is null, this function will return current_time.
But we want to analyze when and why queue_update_finish_time is null, what can we check from clickhouse system table or log ? We've checked the error log and zookeeper log, but it seems there is no infomation about this problem.
Thanks.

apache-zookeeper

来源：https://stackoverflow.com/questions/74065217/how-to-analyze-why-replicasmaxabsolutedelay-return-current-time-problem

1条答案

按热度按时间

y4ekin9u1#

通常，这意味着表不同步，进入只读模式。
导致只读模式的原因可能是：

一些zookeeper同步问题
磁盘存储空间不足
过时的元数据（复制副本无法应用某些更改）
...

要获得更多信息，请查看system.replicas表：

SELECT is_readonly, zookeeper_exception 
FROM system.replicas 
WHERE table = 'appsflyer_events_hist'

在大多数情况下，它有助于运行：

SYSTEM RESTART REPLICA db.table

https://clickhouse.com/docs/en/sql-reference/statements/system/#restart-replica

赞(0）回复(0）举报 2022-12-09

我来回答

Zookeeper 如何分析ReplicasMaxAbsoluteDelay为什么返回当前时间问题？

1条答案

相关问题

热门标签

最新问答