You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@iotdb.apache.org by "Bo Zhang (Jira)" <ji...@apache.org> on 2022/02/28 02:26:00 UTC
[jira] [Commented] (IOTDB-2605) Some issue/bug for new_sync
[ https://issues.apache.org/jira/browse/IOTDB-2605?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17498686#comment-17498686 ]
Bo Zhang commented on IOTDB-2605:
---------------------------------
According to the discussion with [~yschengzi] , the first 3 issues are not problem.
The fourth point will be enhanced in new version of new_snyc.
> Some issue/bug for new_sync
> ---------------------------
>
> Key: IOTDB-2605
> URL: https://issues.apache.org/jira/browse/IOTDB-2605
> Project: Apache IoTDB
> Issue Type: Bug
> Reporter: Bo Zhang
> Assignee: yusicheng
> Priority: Major
> Fix For: 0.13.0
>
> Attachments: image-2022-02-24-17-37-41-058.png, image-2022-02-24-17-50-39-645.png, image-2022-02-24-18-59-43-575.png
>
>
> In branch: new_sync
> 1) In org.apache.iotdb.db.engine.storagegroup.TsFileProcessor#registerSyncDataCollector
> !image-2022-02-24-17-37-41-058.png!
> syncDataCollector 是允许不为空的时直接覆盖吗?。
> 前面三处使用 syncDataCollector 的时候,好像有的加了锁 flushQueryLock ,有的没加,例如 endFile()。
>
> 2)In org.apache.iotdb.db.metadata.mtree.MTree#getTimeseriesAsPlan
> !image-2022-02-24-17-50-39-645.png!
> 这里直接类型强转,是否安全?
>
> 3)In org.apache.iotdb.db.metadata.sync.MetadataSyncManager#registerSyncTask
> 此函数是否会被并发调用,是否需要加锁 保护?
>
> 4) org.apache.iotdb.db.newsync.sender.pipe.TsFilePipe
> !image-2022-02-24-18-59-43-575.png!
> 此文件里,既然 pipeDataDeque 是 BlockingDeque, 为什么使用它 前还要对其加锁? 直接使用 BlockingDeque 的阻塞接口(比如 put、take)不行吗?
>
>
>
>
>
>
>
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)