You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@iotdb.apache.org by "Tan xinyu (Jira)" <ji...@apache.org> on 2020/09/10 08:28:00 UTC

[jira] [Commented] (IOTDB-880) [Distributed] Better concurrecy control in RaftLogManager

    [ https://issues.apache.org/jira/browse/IOTDB-880?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17193458#comment-17193458 ] 

Tan xinyu commented on IOTDB-880:
---------------------------------

Hi, I designed RaftLogManager six months ago with reference to etcd's logmanager. Since the etcd mainly solves the concurrency problem through goroutine's communication by channel rather than locking, I have caused security problems related to lock concurrency after porting it over. Therefore, I intend to improve this module after my current work is completed~

> [Distributed] Better concurrecy control in RaftLogManager
> ---------------------------------------------------------
>
>                 Key: IOTDB-880
>                 URL: https://issues.apache.org/jira/browse/IOTDB-880
>             Project: Apache IoTDB
>          Issue Type: Improvement
>          Components: Core/Cluster
>            Reporter: Tian Jiang
>            Priority: Major
>              Labels: concurrency, raft-log-management
>
> Concurrency control is very primitive in the current version RaftLogManager, especially when there are log capacity checks and log removals, and concurrent requests like "getTerm()" may fail and we must retry. 
> This is not very elegant and it is better to utilize locks to guarantee concurrent safety.



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