changluyi

Results 39 comments of changluyi

I encounter the same issue and committed a pr , but no one reply me https://github.com/brancz/kube-rbac-proxy/pull/282

+1 发现还没解决

+1 发现还没解决

![image](https://github.com/kubeovn/kube-ovn/assets/47097611/cf80e6ba-37b5-4ffe-bfd4-06df16efe385) 看起来像tunnel id给用完了。如果是geneve的话,应该是2 的 16次方个 ![image](https://github.com/kubeovn/kube-ovn/assets/47097611/0d2126f6-18f9-4673-ae70-b1bbe58de25e)

![image](https://github.com/user-attachments/assets/8b4399a9-a73e-43cb-8fa0-21defbf70aca) 目前追溯到是inactivityProbe开了后,这个锁hang住了

目前追溯到rpclient的问题,在这两个位置都能hang住,导致rpcMutex没有释放,尝试更新rpc库没有解决,目前提issue给libovsdb社区。 ![image](https://github.com/user-attachments/assets/3bcfc326-21ee-42f2-ae0d-ef99b9438268) ![image](https://github.com/user-attachments/assets/7c57d0bb-e811-4646-ad81-728501580962) https://github.com/ovn-org/libovsdb/issues/391 看到这个pr 把echo 写成异步的,似乎能解决,但没弄明白当时为啥没合并 https://github.com/cenkalti/rpc2/pull/38 https://github.com/ovn-org/libovsdb/pull/368

”是否应该在ip不足的情况下才去找备选子网” 现在就是这种逻辑吧。 如果你不指定固定ip,就会优先找第一个子网,第一个子网的可用ip置0后,就会找第二个。

> 这个问题好像之前我们修改过,当时的处理策略是, kube-ovn 这边不要处理 nmcli 管理的网卡 是的 打算做一个判断 team网卡别unmanage

![企业微信截图_1729061364984](https://github.com/user-attachments/assets/e2970c1d-08cf-4cd0-9fc0-32f5f1c1cf7f) 手动配置unmanage直接触发网卡down掉

the reason is that , when vm is stopped , the lsp is not deleted yet. then the flow related is also existed ![企业微信截图_17180962966366](https://github.com/kubeovn/kube-ovn/assets/47097611/0ba85294-61d5-4c8e-ae29-676929fe9db8) when other pod send arp request,...