Error Message Is Displayed During TensorFlow 2.6.5 Installation on a Working Node Running EulerOS

Symptom

The error message "Failed to connect to the host via ssh: Shared connection to XX closed" is displayed when TensorFlow 2.6.5 is installed on a working node running EulerOS.

Cause Analysis

The SSH connection session timeout interval is configured on the host. If the time used for task deployment exceeds the configured interval, an error is reported.

Solution

Change the value of ClientAliveInterval in the /etc/ssh/sshd_config file to 1800 (the timeout interval is 30 minutes) and run the following command to restart the sshd service:

systemctl restart sshd