You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "ShuangQi Xia (Jira)" <ji...@apache.org> on 2022/02/08 11:56:00 UTC

[jira] [Commented] (HBASE-26706) The replication batch size is too large and cannot be written to the standby cluster

    [ https://issues.apache.org/jira/browse/HBASE-26706?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17488800#comment-17488800 ] 

ShuangQi Xia commented on HBASE-26706:
--------------------------------------

I think this is diffrent from replication.source.size.capacity , the  replication.source.size.capacity parameter controls all table's size ,but this issue need table's control

> The replication batch size is too large and cannot be written to the standby cluster
> ------------------------------------------------------------------------------------
>
>                 Key: HBASE-26706
>                 URL: https://issues.apache.org/jira/browse/HBASE-26706
>             Project: HBase
>          Issue Type: Bug
>          Components: Replication
>    Affects Versions: 1.3.1, 2.2.3
>            Reporter: ShuangQi Xia
>            Assignee: ShuangQi Xia
>            Priority: Major
>         Attachments: image-2022-01-25-19-59-39-118.png
>
>
> Recently, we met a replication failure and it cannot be recovered. The main reason is that the number of data records read from the wal during synchronization is not limited. In addition, the batch size is not limited when data is written to the peer cluster. As a result, the replication will be failed because of writing timeout. It is recommended that the number of records to be written  with a controllable range, for example, 2000 records, and can be modified through configuration.
> Here are some log stacks
>  !image-2022-01-25-19-59-39-118.png! 
>  



--
This message was sent by Atlassian Jira
(v8.20.1#820001)