You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@iotdb.apache.org by "Houliang Qi (Jira)" <ji...@apache.org> on 2021/01/04 12:04:00 UTC

[jira] [Assigned] (IOTDB-1099) Optimize insertablets logic in cluster module

     [ https://issues.apache.org/jira/browse/IOTDB-1099?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Houliang Qi reassigned IOTDB-1099:
----------------------------------

    Assignee: Houliang Qi

> Optimize insertablets logic in cluster module
> ---------------------------------------------
>
>                 Key: IOTDB-1099
>                 URL: https://issues.apache.org/jira/browse/IOTDB-1099
>             Project: Apache IoTDB
>          Issue Type: Improvement
>          Components: Core/Cluster
>            Reporter: Xinyu Tan
>            Assignee: Houliang Qi
>            Priority: Major
>
> Currently, cluster module will process insertablets rpc by spliting it to multiple insertablet statements one by one, and each of them will take up a raft log,  which need to complete the RaftLogManager's lock and be replicated to followers in one raft log.
> In some user cases, a inserttablets will have 160 tablets, so this single client insertablets rpc will take up 160 raft logs containing insertabletPlan to replicate, which maybe optimized to one raft log containing insertabletsPlan if these devices all belong to same data group; 
> Therefore, we  can process one insertablets rpc by grouping them to several raft logs containing insertabletsPlan according to these belonged data groups, just like the createMultiTimeSeriesPlan in current implementation.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)