Node Connection Failure

Symptom

The analysis task fails. An error is reported indicating the node connection fails.

Possible Cause

• In a multi-node scenario, the times of some nodes may not be calibrated.
• The node where the tool is installed may have a firewall. As a result, the tool cannot communicate with the collected node through a port.

Troubleshooting Procedure

• If the times of some nodes are not calibrated, perform the following operations:

1. Calibrate the time of each node.

   Use the NTP server to synchronize the times, ensuring the consistency between the system time and the standard time.

2. When checking the time of a node, note the time inconsistency caused by the time zones.

   For example, if that the time zone of the node on which the tool is installed is China Standard Time (CST) and that of the node to be analyzed is Eastern Daylight Time (EDT), you can change EDT to CST.

   1 2	mv /etc/localtime /etc/localtime.bak ln -s /usr/share/zoneinfo/Asia/Shanghai /etc/localtime

3. Restart the service after the calibration.
   • root user

     1 2	systemctl restart gunicorn_framework systemctl restart gunicorn_plugin

   • Common user

     1 2	bash /home/test01/DevKit/tools/service_management.sh restart gunicorn_framework bash /home/test01/DevKit/tools/service_management.sh restart gunicorn_plugin

     

     Replace the example installation directory /home/test01 with the actual one.

• If the node connection is blocked by the firewall, perform the following operations:
  • The default communication port is 50051 for the tool installation. If the port is changed during the installation, you can query the value of GRPC_LISTEN_PORT in the /install_path/DevKit/config/install.ini file on the active node and set the firewall to enable access to the communication port.
    

    /install_path/: DevKit installation directory. Replace it with the actual directory.

  • Disable the firewall of the node where the tool is installed. (Common users do not have the permission to perform operations on the firewall.)

    1	systemctl stop firewalld