Hi,
Is it possible, to install CoreOS using ignition, which will rename default interface names to my own, like “production.oam” or “test.backup” - I mean Network Manager names. Interface names as coming from kernel, they can stay, but it would be nice to be able to rename them too.
Here is what I mean, but all this had to be done manually (coreos-installer --copy-network):
[root@ves-sd ~]# ocssh ves-oce4.xxx.okd nmcli conn show
Warning: Permanently added ‘ves-oce4.xxx.okd,xxx’ (ECDSA) to the list of known hosts.
NAME UUID TYPE DEVICE
ves.ipnet.oam bc1e3fac-79b4-4d5e-887a-5b7226dadcef bond bond0
ves.ipnet.backup c1196e6d-75ae-4722-bfa4-62400c77f4d2 vlan bond0.2311
ves.ipnet.db-replication c786b5fb-e2ff-4513-be5f-0e8755ac7035 vlan bond0.2310
ves.ipnet.k8s-to-lb 632c2928-770c-40a4-973a-f70204acd0de vlan bond0.2314
ves.ipnet.sig-int 9f9e49d8-dcf3-4e54-9b8f-c6d25f90a841 vlan bond0.2312
bond0-member1 3f3459fe-5e38-405c-8e72-ba7681731ef0 ethernet ens10f0np0
bond0-member2 4f536044-bad5-4b9a-9960-2d790976d4d6 ethernet ens1f0
Wired connection 1 2690f871-8655-32f3-b77f-c73752ae9ed6 ethernet –
Wired connection 2 ffd85e98-c53b-3b33-b65b-5ef4e53058bd ethernet –
To achieve this, I install CoreOS LIVE image, then I need to access that, enable ssh, set root password, allow root to login (for a while), and access that with another machine to be able to put this:
sudo sed -i 's/^PasswordAuthentication.*/PasswordAuthentication yes/' /etc/ssh/sshd_config; sudo sed -i 's/^PermitRootLogin.*/PermitRootLogin yes/' /etc/ssh/sshd_config; echo 'root:pol1cl' | sudo chpasswd; sudo systemctl restart sshd
nmcli -g UUID con show | while read line; do nmcli con delete uuid $line; done
nmcli con add type bond con-name ves.ipnet.oam ifname bond0 bond.options "mode=active-backup,arp_interval=2000,arp_ip_target=172.20.197.1" ipv4.addresses 172.20.197.12/24 ipv4.gateway 172.20.197.1 ipv4.dns "172.20.197.77,172.20.197.78" ipv4.dns-search "xxx.okd" ipv4.method manual 802-3-ethernet.mtu 9000 ipv6.method disabled
nmcli con add type ethernet con-name bond0-member1 ifname ens10f0np0 master ves.ipnet.oam 802-3-ethernet.mtu 9000
nmcli con add type ethernet con-name bond0-member2 ifname ens1f0 master ves.ipnet.oam 802-3-ethernet.mtu 9000
nmcli con up ves.ipnet.oam; sleep 1
nmcli con add type vlan con-name ves.ipnet.backup dev bond0 id 2311 ip4 10.20.48.12/24 802-3-ethernet.mtu 9000 ipv6.method disabled; nmcli conn up ves.ipnet.backup
nmcli con add type vlan con-name ves.ipnet.db-replication dev bond0 id 2310 ip4 172.20.199.12/24 802-3-ethernet.mtu 9000 ipv6.method disabled; nmcli conn up ves.ipnet.db-replication
nmcli con add type vlan con-name ves.ipnet.k8s-to-lb dev bond0 id 2314 ip4 10.20.55.12/24 802-3-ethernet.mtu 9000 ipv6.method disabled; nmcli conn up ves.ipnet.k8s-to-lb
nmcli con add type vlan con-name ves.ipnet.sig-int dev bond0 id 2312 ip4 10.20.51.12/24 802-3-ethernet.mtu 9000 ipv6.method disabled; nmcli conn up ves.ipnet.sig-int
#coreos-installer install --copy-network --insecure-ignition --ignition-url=http://172.20.197.72:8080/xxx/worker1.ign /dev/sda
But this is very time consuming, especially if I need to do it on 100 nodes.
Unfortunately, I install this Openshift on baremetal HP ProLiant machines, each of them boots few minutes.
Any way to automate it (automatic boot to CoreOS LIVE image), set network and create cluster?
I cannot use automatically generated and customized ISO files, because due to HP issue, such ISO files cannot recognize network (cannot initialize it), so the only way is to boot into LIVE image, setup network and then install CoreOS on top of this.