preface

Recently, I am using BCS service of Huawei Cloud, because I am using Huawei chain service. Because it is still in beta version, its functions are not perfect. Although it has built-in cluster deployment function, it is quite troublesome to encounter some problems.

Problems and Solutions

0/1 nodes are available: 1 node(s) didn't match pod affinity/anti-affinity, 1 node(s) didn't satisfy existing pods anti-affinity rules.

Scheduling failure Major CCE service clusterName=cluster-bcs-xzzv; clusterID=7ca30461-b4c8-11eb-bfc3-0255ac1022c9; kind=Pod; namespace=kube-system; name=everest-csi-controller-54b67f46bd-trs5h; uid=02ba688f-28c9-4d6e-938e-c7ca0af7c026 0/1 nodes are available: 1 node(s) didn't match pod affinity/anti-affinity, 1 node(s) didn't satisfy existing Pods anti-affinity rules. 2021/05/15 11:50:51 GMT+08:00 16sec clusterName=cluster-bcs-xzzv; clusterID=7ca30461-b4c8-11eb-bfc3-0255ac1022c9; kind=Pod; namespace=kube-system; name=coredns-8687fc4977-8wvb9; uid=3c174f7d-6492-49f9-a8b1-6a056fc141dc 0/1 nodes are available: 1 node(s) didn't match pod affinity/anti-affinity, 1 node(s) didn't satisfy existing pods anti-affinity rules. 2021/05/15 11:50:51 GMT+08:00Copy the code

You can see the alarm informationeverest-csi-controllerandcorednsThe two POD schedules failed. See the alarm details below

0/1 nodes are available: 1 node(s) didn't match pod affinity/anti-affinity, 1 node(s) didn't satisfy existing pods anti-affinity rules
Copy the code

This means that these two Pods do not conform to the POD affinity scheduling policy. If you want to know more, you can see the k8S concept of affinity reference blog.csdn.net/qq_38093301…

Ok through the blog we can also roughly the problem, we open huawei cloud CCE cluster management, view the container group POD

Cloud Container Engine -> Workload -> Container group POD select all namespacesSo you can see that these two pods actually failed the schedule so let’s get rid of these two

You can see that the chaining code has been successfully deployed to both nodes simultaneously

Other problems

Cloud Container Engine -> Workloads -> Stateless Loads Click the load nameeverest-csi-controller After deletion will continue to dispatch, and the affinity policy can not be modified (K8S system POD), then I have no way, anyway, the chain code at the same time deployment problem I do not bother, some students know the solution welcome to leave comments, other deployment problems can also find bloggers to discuss (*￣) ￣)

Add nodes to CCE cluster