That’s great. I’m going to do it on OpenBSD with relayd if possible.
im just glad im using ydotool to auto type stuff into the VM virt-viewer consoles…
aka cheating?
What is that? autocomplete?
no its scriptable typing or key board press emulation ( guess is the word for it)
oh so like you press F and it runs something? like a macro?
i have it in bash script…
i run script and there are sleeps in script so i have time to click on each vm viewer console window after it types and hits enter
sample
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/bootstrap.ign /dev/sda --ignition-hash=sha512-${bootstrap} && sleep 1 && shutdown -r " && sudo ydotool key 28:1 28:0
sleep 1
echo worker
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/worker.ign /dev/sda --ignition-hash=sha512-${worker} && sleep 200 && shutdown -r " && sudo ydotool key 28:1 28:0
echo worker
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/worker.ign /dev/sda --ignition-hash=sha512-${worker} && sleep 200 && shutdown -r " && sudo ydotool key 28:1 28:0
sleep 1
echo master
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/master.ign /dev/sda --ignition-hash=sha512-${master} && sleep 8 && shutdown -r " && sudo ydotool key 28:1 28:0
echo master
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/master.ign /dev/sda --ignition-hash=sha512-${master} && sleep 21 && shutdown -r " && sudo ydotool key 28:1 28:0
echo master
sleep 1
sleep 4 && sudo ydotool type "bash" && sudo ydotool key 28:1 28:0 && sudo ydotool type "sudo coreos-installer install --ignition-url=http://f35-manager.okd.lab:8080/okd4/master.ign /dev/sda --ignition-hash=sha512-${master} && sleep 28 && shutdown -r " && sudo ydotool key 28:1 28:0
echo
echo
echo done
im only typing ‘bash’ and hitting enter so i see that the PS1 updates with the hostname given by dhcp…fyi
Oh I see because you have to enter that into the consoles of each vm? Surely there’s a way to bootstrap the CoreOS instances without console intervention?
theres a section for pxe that i havent read…
might be possible to do it that way
Finally 2nd success
NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE
authentication 4.10.0-0.okd-2022-03-07-131213 True False False 42m
baremetal 4.10.0-0.okd-2022-03-07-131213 True False False 57m
cloud-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 61m
cloud-credential 4.10.0-0.okd-2022-03-07-131213 True False False 62m
cluster-autoscaler 4.10.0-0.okd-2022-03-07-131213 True False False 56m
config-operator 4.10.0-0.okd-2022-03-07-131213 True False False 58m
console 4.10.0-0.okd-2022-03-07-131213 True False False 46m
csi-snapshot-controller 4.10.0-0.okd-2022-03-07-131213 True False False 57m
dns 4.10.0-0.okd-2022-03-07-131213 True False False 56m
etcd 4.10.0-0.okd-2022-03-07-131213 True False False 55m
image-registry 4.10.0-0.okd-2022-03-07-131213 True False False 48m
ingress 4.10.0-0.okd-2022-03-07-131213 True False False 48m
insights 4.10.0-0.okd-2022-03-07-131213 True False False 51m
kube-apiserver 4.10.0-0.okd-2022-03-07-131213 True False False 48m
kube-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 53m
kube-scheduler 4.10.0-0.okd-2022-03-07-131213 True False False 53m
kube-storage-version-migrator 4.10.0-0.okd-2022-03-07-131213 True False False 57m
machine-api 4.10.0-0.okd-2022-03-07-131213 True False False 57m
machine-approver 4.10.0-0.okd-2022-03-07-131213 True False False 57m
machine-config 4.10.0-0.okd-2022-03-07-131213 True False False 55m
marketplace 4.10.0-0.okd-2022-03-07-131213 True False False 56m
monitoring 4.10.0-0.okd-2022-03-07-131213 True False False 42m
network 4.10.0-0.okd-2022-03-07-131213 True False False 58m
node-tuning 4.10.0-0.okd-2022-03-07-131213 True False False 57m
openshift-apiserver 4.10.0-0.okd-2022-03-07-131213 True False False 51m
openshift-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 51m
openshift-samples 4.10.0-0.okd-2022-03-07-131213 True False False 51m
operator-lifecycle-manager 4.10.0-0.okd-2022-03-07-131213 True False False 57m
operator-lifecycle-manager-catalog 4.10.0-0.okd-2022-03-07-131213 True False False 57m
operator-lifecycle-manager-packageserver 4.10.0-0.okd-2022-03-07-131213 True False False 51m
service-ca 4.10.0-0.okd-2022-03-07-131213 True False False 57m
storage 4.10.0-0.okd-2022-03-07-131213 True False False 58m
the machine-config operater actually finished quickly
the only things i changed since the last try was adding more RAM to my registry mirror vm, and adding the other pull-secrets that i have when i setup the registry mirror
OK got my third success.
I believe throwing some more cores at it helped this time.
now; the masters have 5, the workers have 4, the bootstrap has 4 and the mirror has 5.
The machine-config part finished in under a minute… so hopefully Ive found the sweet spot of cores and memory for these VMs.
NAME VERSION AVAILABLE PROGRESSING DEGRADED SINCE MESSAGE
authentication 4.10.0-0.okd-2022-03-07-131213 True False False 31m
baremetal 4.10.0-0.okd-2022-03-07-131213 True False False 46m
cloud-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 50m
cloud-credential 4.10.0-0.okd-2022-03-07-131213 True False False 50m
cluster-autoscaler 4.10.0-0.okd-2022-03-07-131213 True False False 47m
config-operator 4.10.0-0.okd-2022-03-07-131213 True False False 47m
console 4.10.0-0.okd-2022-03-07-131213 True False False 36m
csi-snapshot-controller 4.10.0-0.okd-2022-03-07-131213 True False False 47m
dns 4.10.0-0.okd-2022-03-07-131213 True False False 46m
etcd 4.10.0-0.okd-2022-03-07-131213 True False False 46m
image-registry 4.10.0-0.okd-2022-03-07-131213 True False False 40m
ingress 4.10.0-0.okd-2022-03-07-131213 True False False 38m
insights 4.10.0-0.okd-2022-03-07-131213 True False False 41m
kube-apiserver 4.10.0-0.okd-2022-03-07-131213 True False False 40m
kube-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 44m
kube-scheduler 4.10.0-0.okd-2022-03-07-131213 True False False 43m
kube-storage-version-migrator 4.10.0-0.okd-2022-03-07-131213 True False False 47m
machine-api 4.10.0-0.okd-2022-03-07-131213 True False False 47m
machine-approver 4.10.0-0.okd-2022-03-07-131213 True False False 47m
machine-config 4.10.0-0.okd-2022-03-07-131213 True False False 47m
marketplace 4.10.0-0.okd-2022-03-07-131213 True False False 46m
monitoring 4.10.0-0.okd-2022-03-07-131213 True False False 33m
network 4.10.0-0.okd-2022-03-07-131213 True False False 48m
node-tuning 4.10.0-0.okd-2022-03-07-131213 True False False 46m
openshift-apiserver 4.10.0-0.okd-2022-03-07-131213 True False False 40m
openshift-controller-manager 4.10.0-0.okd-2022-03-07-131213 True False False 41m
openshift-samples 4.10.0-0.okd-2022-03-07-131213 True False False 41m
operator-lifecycle-manager 4.10.0-0.okd-2022-03-07-131213 True False False 47m
operator-lifecycle-manager-catalog 4.10.0-0.okd-2022-03-07-131213 True False False 46m
operator-lifecycle-manager-packageserver 4.10.0-0.okd-2022-03-07-131213 True False False 42m
service-ca 4.10.0-0.okd-2022-03-07-131213 True False False 47m
storage 4.10.0-0.okd-2022-03-07-131213 True False False 47m
Ordered a replacement fan for my NUC8…
got it replaced and so far… i havent heard that awful noise again yet…
I think the OKD install is reaching out to something on the internet… and thats why the install is not 100% every try. (damn machine-config clusteroperator)
I might muck up the default routing to see what happens if the VMs cant get to the internet at all.
guess i am going to double the ram on the box ive been using for my OKD testing… that way i can try to figure out why its only successfully installed 4 out of like 3 dozen attempts.
I moved my openshift cluster off oVirt and onto baremetal. Then I tried to move from a USG Pro to a UDM SE so I don’t need a controller vm. Didn’t realize how bad the unifi stocking issues were and have had a mixed infrastructure and a half-done migration sitting in limbo for months.
Now my UDM SE should be here on Monday which means I can take the beefier servers off oVirt and add them to the OpenShift cluster.
It probably wouldn’t matter, but I was too scared to try anything new while using this mixed infra. I have some ‘critical’ services up like nextcloud and vaultwarden, but now I can go do some more experimenting.
ive got the bare metal install down now…
Really minimal input needed from me to install it since after about the first 7-8 attempts i scripted it.
Which is how ive done dozens of attempts in my testing…
my problem is… after its going I know absolutely nothing on how to use it or admin it… but i can install it
The assisted installer is GA now (I think) and it really makes things simpler.
Also, helm is your friend for installing kube stuff if you’re new-ish to it.
again…
i know of basic podman and docker stuff manually, but the world of k8/open shift… im still in the clueless phase.