[
https://issues.apache.org/jira/browse/IOTDB-4066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17634633#comment-17634633
]
Yongzao Dan commented on IOTDB-4066:
------------------------------------
We'll solve this problem by double keyword sort algorithm:
# Choose the DataNode that has the least number of DataRegion.
# Choose the DataNode that has the maximum free disk space.
https://apache-iotdb.feishu.cn/docs/doccnHuDkDD6RiwLjaej3hkA9Vb#
> 【1】The allocation of regions and data needs to consider the available disk
> space of the node
> --------------------------------------------------------------------------------------------
>
> Key: IOTDB-4066
> URL: https://issues.apache.org/jira/browse/IOTDB-4066
> Project: Apache IoTDB
> Issue Type: Improvement
> Components: mpp-cluster
> Reporter: 刘珍
> Assignee: 陈哲涵
> Priority: Minor
> Attachments: image-2022-08-08-18-13-49-091.png,
> image-2022-08-08-18-15-39-357.png
>
>
> region 和数据的分配，需要考虑节点磁盘空间：
> 如下启动测试前，各节点的磁盘空间信息，ip4的可用磁盘空间少：
> !image-2022-08-08-18-13-49-091.png!
> 3副本*3 dataregion
> ，7datanode，会多出来2个副本，其中的1个副本，分配给了可用磁盘少的ip4，写入数据一段时间后ip4发生磁盘空间不足：
> !image-2022-08-08-18-15-39-357.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
