[IOTDB-4772] error would occurs when starting datanode 2 times and all nodes seems normal with show cluster

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.14.0-SNAPSHOT
• Component/s: Tools/Others
• Labels: pull-request-available
• Sprint: 2022-11-Cluster
• Assignee: yanze chen Assign to me
• Reporter: changxue
• Created: 27/Oct/22 03:26
• Updated: 18/Nov/22 08:11
• Resolved: 14/Nov/22 05:51

Description

error would occurs when starting datanode 2 times and all nodes  seems normal with show cluster

reproduction:

1. start a 3C3D iotdb cluster successfully
2. run start-datanode.sh on 1 node without stopping it

 

actual result:

1. show cluster  shows all confignode and datanode is running
2. but error occurs when create timeseries

 

建议：

1. 在start-confignode.sh 和 start-datanode.sh 脚本中，应该先判断是否已经有启动的服务，如果有，先停止，然后再启动。
2. 咱们的stop-datanode.sh 内停止服务的方式不对，现状是使用ps然后查找datanode字样。一般通用做法是，在启动时将PID记录在pid文件中，停止时，读取并kill。目前的做法容易误停止其他服务（或者有多个时直接报错）。

Issue Links

• is related to: IOTDB-3732 [same folder but different rpc port]- "io task meets error" is not handled correctly
• relates to: IOTDB-4402 Avoid multiple IoTDB processes accessing same data dir
• links to: GitHub Pull Request #7562