Network congestion caused by docker Network failure

failure scenarios

Today, I received feedback from developers on the project. After installing docker, a customer has a VM. When starting some docker Kubernetes Engine, he found that these Kubernetes Engine could not start normally. View system log via tail -f /var/log/messages

It is found that the system has been reporting errors, mainly bro: port2 (veth0) entered blocking state or bro: port1 (vxlan0) entered blocking state, and the NIC is in blocking state. See the following figure for specific error reports:

Use systemctl status docker -l to check the docker running status, and find that docker is running normally, but there are errors printed in the status information, which are all network errors ~

Troubleshooting and Handling

The above docker running status is normal, indicating that the installation and deployment of docker should be no problem. From the error report, it should be a problem with the host network. After investigation, it is found that the system's selinux is in the enforcing status.

Temporarily shut down selinux restart docker found that the network returned to normal ~

The main screening steps are as follows:

(1) View the current selinux status of the system

getenforce

(2) Close selinux

#temporary shutdown setenforce 0#permanent shutdown sed -i s#SELINUX=enforcing#SELINUX=disabled#g /etc/sysconfig/selinux

(3) Check the system firewall status

#view status systemctl status firewalld#close firewall (on demand) systemctl stop firewalld

(4) Restart docker service

systemctl restart docker