RHEL8 does not ship a XEN-capable network installer anymore. This is why it is now best to simply grab the latest cloud image.
cd /data_local/ wget https://cloud.centos.org/centos/8-stream/x86_64/images/CentOS-Stream-GenericCloud-8-latest.x86_64.qcow2 du -sh CentOS-Stream-GenericCloud-8-latest.x86_64.qcow2 # 1.9G ln -s CentOS-Stream-GenericCloud-8-latest.x86_64.qcow2 centos8.qcow2
guest=centos8 mkdir -p ~/guests/$guest/ cd ~/guests/$guest/
what is the max amount of cores you can share with the guest?
grep ^proc /proc/cpuinfo | tail -1
booting the cloud image with rescue mode so you can reset root password
vi $guest
kernel = "/data/kernels/5.2.21.domureiser4.vmlinuz" root = "/dev/xvda1 ro console=hvc0 net.ifnames=0 biosdevname=0 mitigations=off" extra = "init=/bin/bash" name = "centos8" vcpus = 3 maxvcpus = 31 memory = 1024 #memory = 7168 disk = ['qcow2:/data_local/centos8.qcow2,xvda,w'] vif = [ 'bridge=guestbr0, vifname=centos8.0' ] type = "pvh"
note we are not trying to enable pvgrub here, as PVH does not support it
get into the rescue mode shell
xl create centos8 -c export PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin mount -o remount,rw / mount -t proc none /proc cat /proc/partitions
you should then proceed similarly to the system preparation but there’s a few differences. first, there are more services to disable.
for sv in \
cloud-config \
cloud-final \
cloud-init-local \
cloud-init \
cloud-config.target \
NetworkManager \
kdump \
; do
# self-verbose
echo disabling $sv
systemctl disable $sv || echo FAILED
echo
done; unset sv
you would also need to grab modules for your custom kernel, if any (e.g. TMEM here)
dhclient -v eth0 ping -c1 opendns.com yum update -y yum install network-scripts wget mlocate -y systemctl list-unit-files | grep network systemctl enable network wget https://lab.nethence.com/nunux/5.2.21.domureiser4.modules.tar.gz tar xzf 5.2.21.domureiser4.modules.tar.gz -C /lib/modules/ rm -f 5.2.21.domureiser4.modules.tar.gz echo tmem > /etc/modules-load.d/tmem.conf
that cloud image also has additional shit to get rid of
ls -l /etc/motd.d/cockpit ls -l /etc/issue.d/cockpit.issue ls -l /etc/sysconfig/network-scripts/ifcfg-* rm -f /etc/motd.d/cockpit rm -f /etc/issue.d/cockpit.issue rm -f /etc/sysconfig/network-scripts/ifcfg-ens3
eventually initialize the root password
#passwd root
–or– push your ssh pubkey there
mkdir /root/.ssh/ cat > /root/.ssh/authorized_keys <<EOF ... EOF chmod 700 /root/ chmod 700 /root/.ssh/ chmod 600 /root/.ssh/authorized_keys
when finished, shutdown the guest
sync halt -p #xl des centos8 reset
and disable the extra init
vi centos8 #extra = "init=/bin/bash" extra = "ip=10.1.0.9::10.1.0.254:255.255.255.0:centos8:eth0:none"
you’re now good to go for acceptance testing
xl create centos8 -c
warning: since we don’t mount and edit the QCOW2 disk from the host, we also need to be careful about the SSH host keys. Those should be removed before generating the template.
eventually make a template out of it
tar cSf /data/templates/centos.qcow2.tar centos.qcow2
https://www.centos.org/download/mirrors/
https://wiki.centos.org/action/show/Sources
https://pagure.io/centos-kernel-rpm/blob/c8/f/SOURCES/kernel-x86_64.config
https://pagure.io/centos-kernel-rpm/blob/c8/f/SOURCES
Bug 1663812 - document method to disable motd message https://bugzilla.redhat.com/show_bug.cgi?id=1663812
Is there a way to permanently remove the cockpit.socket message? https://forums.centos.org/viewtopic.php?t=72303
Strange message on ssh login https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org/thread/AQH3F2KY5RUP2RC7EDDXRTMW3AVP2DZJ/
CHAPTER 3. MANAGING KERNEL MODULES https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/managing_monitoring_and_updating_the_kernel/managing-kernel-modules_managing-monitoring-and-updating-the-kernel
https://anaconda-installer.readthedocs.io/en/latest/boot-options.html