However, when there is two ips starting with 10, this script fails and leave a unclean environment. I try to use scripts/github_runner/clean_cri_runner.sh and scripts/clean_fcctr.sh to clean and reinstall, I enconter error failed calling webhook "ipaddresspoolvalidationwebhook.metallb.io". To restore a clean environment, I have to execute these two cleaning scripts and reboot machine.
Therefore, there are actually two bugs: one is the inability to handle the situation where there are two IPs starting with 10, and the other is that the clean script is ineffective in this situation.
To Reproduce
Deploy in a multi-node environment, the master node has two network interfaces with IPs starting with '10'.
Machine: x86_64
OS: ubuntu20.04 LTS server
Software: stock-only
Expected behavior
Inform the user to select which network interface card to use and provide a script that can properly clean the environment.
Describe the bug When deploy kubernetes, create_multinode_cluster.go will obtain ip starting with 10 and pass it as a parameter for kubeadm.
However, when there is two ips starting with 10, this script fails and leave a unclean environment. I try to use
scripts/github_runner/clean_cri_runner.shandscripts/clean_fcctr.shto clean and reinstall, I enconter errorfailed calling webhook "ipaddresspoolvalidationwebhook.metallb.io". To restore a clean environment, I have to execute these two cleaning scripts and reboot machine. Therefore, there are actually two bugs: one is the inability to handle the situation where there are two IPs starting with 10, and the other is that the clean script is ineffective in this situation. To Reproduce Deploy in a multi-node environment, the master node has two network interfaces with IPs starting with '10'. Machine: x86_64 OS: ubuntu20.04 LTS server Software: stock-only Expected behavior Inform the user to select which network interface card to use and provide a script that can properly clean the environment.