kube-ovn icon indicating copy to clipboard operation
kube-ovn copied to clipboard

A Bridge between SDN and Cloud Native (Project under CNCF)

Results 274 kube-ovn issues
Sort by recently updated
recently updated
newest added

# Expected Behavior Two staticIP pod can ping each other # Actual Behavior Two pods can not ping each other # Steps to Reproduce the Problem 1.create statefulset-xyz 2.create statfulset-uvw...

# Expected Behavior vm still works well when kube-ovn-controller set keep-vm-ip = true, and reboot all vm # Actual Behavior # Steps to Reproduce the Problem use kube-ovn 1.10.0 1....

k3s有一个svclb-traefik,会添加一条主机80 443端口的流量DNAT到svclb-traefik Pod的iptables规则: ![image](https://user-images.githubusercontent.com/35413234/178389366-ad9579e5-0a0b-4662-b4d2-1b56ab72e5cb.png) 1.10.2版本未生成该iptables规则: ![image](https://user-images.githubusercontent.com/35413234/178388555-44fd09d1-b4f4-4f58-809b-6ed39bdb1ff7.png) 导致双网络接口一个可以通,一个不通: ![image](https://user-images.githubusercontent.com/35413234/178388711-a74113e4-8c81-4480-987f-e609966d6191.png) 而1.9.5版本没有该问题: ![image](https://user-images.githubusercontent.com/35413234/178389178-040ffb3f-d75b-4602-9921-96da6d5a2838.png) 此外,1.10.2版本中ovn-central组件一直报错: ![image](https://user-images.githubusercontent.com/35413234/178389570-d7e3ab83-742b-481a-9042-a7e7c8d72d8a.png) 1.9.5版本无此问题。 感谢。

We only have doc about chaining mode with Cilium now. But we need a guide to show what capabilities we have and how to show them.

docs

# Expected Behavior pod内部获取到的web请求的源ip是外部请求发起的IP地址。 # Actual Behavior pod内部获取到的web请求的源ip是node节点在join子网的IP地址。 # Steps to Reproduce the Problem 1. 集群外node访问nodeport 的ip 2. pod内部打印remoteAddr # Additional Info - Kubernetes version: **Output of `kubectl version`:** ```...

# Expected Behavior 解决偶现的问题 # Actual Behavior # Steps to Reproduce the Problem 1. 重复创建nat-gw-pod,创建和删除 fip 有概率出现 # ping fip 不通时的情况 ![image](https://user-images.githubusercontent.com/7981158/174558780-679930f6-c47b-40b0-b7df-91a33b3451b5.png) # ping fip 正常时的情况 ![image](https://user-images.githubusercontent.com/7981158/174558929-55aa3fa5-4524-4b7a-ad6a-72209ce2f697.png) 对比发现 少了net1网卡 ttl...

目前像deployment这类workload资源,只有ip_pool配置,现在有一个场景需要mac地址也可以像ip_pool一样有一个可分配的池,不知有有没有什么解决方案? apiVersion: apps/v1 kind: Deployment metadata: namespace: ovn-test name: starter-backend labels: app: starter-backend spec: replicas: 2 selector: matchLabels: app: starter-backend template: metadata: labels: app: starter-backend annotations: ovn.kubernetes.io/ip_pool: 10.16.0.15,10.16.0.16,10.16.0.17 #如何实现mac_pool spec:...

feature

# Expected Behavior # Actual Behavior # Steps to Reproduce the Problem 通过nsenter进到容器的namespace中,执行arping 命令,返回执行结果为成功 ![image](https://user-images.githubusercontent.com/11764870/173996060-3dc68eae-dd6b-4cd8-a87a-081834403cb8.png) 而直接进入容器,执行arping命令,返回执行结果为失败 ![5c77f9f34ee1ed5bb2e368a7f951e62](https://user-images.githubusercontent.com/11764870/173996146-9af21522-db11-4dfb-957b-2fba0f915017.png) 这导致无法成功创建EIP,ovn-controller报错如下: ![image](https://user-images.githubusercontent.com/11764870/173996309-5f0a9e0d-20f2-44b7-9404-7d405125b29c.png) # Additional Info - Kubernetes version: **Output of `kubectl version`:** ```...

# Expected Behavior # Actual Behavior # Steps to Reproduce the Problem 1. 2. 3. # Additional Info - Kubernetes version: **Output of `kubectl version`:** ``` (paste your output here)...

bug

# Expected Behavior 当存在大量 POD 时,占用内存 18G 左右,不知道占用是否正常。 删除 POD 后,占用的内存应该释放。 # Actual Behavior kube-ovn-controller 在大量 POD (1W)场景,占用内存 18G 左右,删除 POD 后,内存未释放。 # Steps to Reproduce the Problem 1. 使用...