currently manual grub changes are needed we'd like to do that in UI, somewhere under Hosts. Optional functionality to update and regenerate grub when host is in Maintenance, then reboot. Additional parameters for e.g. GPU passthrough might be needed. it needs to support node as well
For reference vintage node is using grub2 - but the grub2 configuration file is not under /boot. In addition vintage Node is _only_ caring about grub.conf - and not about the additional files /etc/sysconfig/grub and others.
If it's only about 4.0, then Node Next will be used, which will work as CentOS/RHEL does.
Just for reference: There is already a tool to add customer kernel arguments to the bootloader config: grubby. https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/System_Administrators_Guide/sec-Making_Persistent_Changes_to_a_GRUB_2_Menu_Using_the_grubby_Tool.html grubby is currently the most likely candidate to be used on Node Next as well for bootloader configuration. It supports grub, grub2 and syslinux and other bootloaders.
grubby should be a good solution for 4.0, we won't have to deal with grub directly
Let me note that oVirt Node Next is now using grubby (as recommended) to modify the grub.cfg file.
Don't you also need to boot the host (I wonder if kexec would suffice?)
Yes, it needs to be rebooted. But if we make it a part of redeploy it's ok, we can do a full "normal" reboot.
Moving from 4.0 alpha to 4.0 beta since 4.0 alpha has been already released and bug is not ON_QA.
changes are in, just waiting on host-deploy respin and spec update
Verified: rhevm-4.0.2.2-0.1.el7ev qemu-kvm-rhev-2.3.0-31.el7_2.20.x86_64 libvirt-client-1.2.17-13.el7_2.5.x86_64 vdsm-4.18.9-1.el7ev.x86_64 Test plan attached to external trackers.
Hm, it seems documentation was not really updated with the new feature. Let me open a new doc bug for that as we need to update screenshots of Cluster dialog and add a bit more details about supportability of each kernel option