Well, ens is missing because with the system-config-kickstart tool, you actually specify your own interface name. I chose eth0 and 1, but I could have just as easily typed eth0 and ens4, which is what CentOS defaults to during both a manual and an automatic install.
I will need to look into the script some more. I am new to kickstart and I haven't had much of a chance to work with postinstall scripts.
Issues with using kickstart to program more than 1 NIC on boot
Re: Issues with using kickstart to program more than 1 NIC on boot
If one can name interfaces, then I would use descriptive names like "data" and "man". No digits.
If the auto-generated eth/ens names depend on hypervisor device model, and names are different, then why the two interfaces do use different models?
If the auto-generated eth/ens names depend on hypervisor device model, and names are different, then why the two interfaces do use different models?
-
- Posts: 2019
- Joined: 2015/02/17 15:14:33
- Location: Bulgaria
- Contact:
Re: Issues with using kickstart to program more than 1 NIC on boot
As per the discussion in RHEL forum , you need to pass the following kernel parameter in the VM's grub line in order to disable the new naming convention.That's what this postinstall script (in my previous message) should do. Then the kernel will name the first device eth0 , second eth1 and so on in order they are defined in the VM's configuration. I haven't tried it , but it's worth trying.
Code: Select all
net.ifnames=0