1. Environment configuration
| k8s version | docker version | installation method | network plugin |
|---|---|---|---|
| 1.19.14 | 18.06.3 | kubeadm+haproxy+keepalived | it's flannel |
2. As shown in the figure below, some nodes appear NotReady during operation.
3. After investigation, it is found that there are three pods that have not been created.
4. View one of the pods through describe
and you can see that the pod has been scheduled to the master2 node. The information in the previous part found that there are stains
5. Continuing from part 4, check the nodes through describe, and found that there are several stains, and they are all stains of k8s
.
##手动修改一下网卡然后重启
vim /etc/sysconfig/network-scripts/ifcfg-ens192
systemctl restart network
7. k8s comes with stains
| taint | analyze |
|---|---|
| node.kubernetes.io/disk-pressure | The pressure on the disk is high, and the storage may be insufficient. Check the disk |
| node.kubernetes.io/out-of-disk | Insufficient disk space, check disk |
| node.kubernetes.io/not-ready | not ready |
| node.kubernetes.io/unreachable | The node cannot be accessed, check the network problem |
| node.kubernetes.io/memory-pressure | The memory pressure is high, check the memory |
| node.kubernetes.io/network-unavailable | The network is not available, check the network |
| node.kubernetes.io/unschedulable | Node cannot be scheduled |