Hi Gianluca, As mentioned in below RFE, EL 8.4 is not enough, you also need Advanced Virtualization 8.4 (libvirt >= 7.0.0) and AFAIK this has not been released upstream: https://bugzilla.redhat.com/show_bug.cgi?id=1933974 So right now you can use 4.6 cluster level only on hosts running on CentOS Stream. Also the error message tells you that all your Up hosts need to provide EL + AV 8.4 features before you upgrade cluster level. But of course you can upgrade them one by one, it's not required to put all hosts into Maintenance before upgrade. Regards, Martin On Tue, Jun 8, 2021 at 3:40 PM Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt; wrote: -- Martin Perina Manager, Software Engineering Red Hat Czech s.r.o.

On Tue, Jun 8, 2021 at 4:10 PM Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt; wrote: Relevant bugs are linked in Blocks field in Bugzilla: *1669178* <https://bugzilla.redhat.com/show_bug.cgi?id=1669178> - [RFE] Q35 SecureBoot - Add ability to preserve variable store certificates *1688177* <https://bugzilla.redhat.com/show_bug.cgi?id=1688177> - [RFE] [SR-IOV] [Tech Preview RHV 4.4.6] Migration should not require downtime as of today *1821199* <https://bugzilla.redhat.com/show_bug.cgi?id=1821199> - HP VM fails to migrate between identical hosts (the same cpu flags) not supporting TSC *1834250* <https://bugzilla.redhat.com/show_bug.cgi?id=1834250> - CPU hotplug on UEFI VM causes VM reboot *1892800* <https://bugzilla.redhat.com/show_bug.cgi?id=1892800> - NVDIMM: VM hangs in wait for launch for more than 7 minutes before VM is up *1906074* <https://bugzilla.redhat.com/show_bug.cgi?id=1906074> - [RFE] Support disks copy between regular and managed block storage domains *1927718* <https://bugzilla.redhat.com/show_bug.cgi?id=1927718> - [RFE] Provide Reset option for VMs *1936163* <https://bugzilla.redhat.com/show_bug.cgi?id=1936163> - Enable bochs-display for UEFI guests by default *1936164* <https://bugzilla.redhat.com/show_bug.cgi?id=1936164> - Enable KVM Software TPM by default *1950752* <https://bugzilla.redhat.com/show_bug.cgi?id=1950752> - [RFE][CBT] redefine only the checkpoint that the backup is taken from and not the entire chain Yes, oVirt Node 4.4.6 is based CentOS Stream: *1907833* <https://bugzilla.redhat.com/show_bug.cgi?id=1907833> - Rebase oVirt Node on CentOS Stream 8 -- Martin Perina Manager, Software Engineering Red Hat Czech s.r.o.

On Tue, Jun 8, 2021 at 6:14 PM Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt; wrote: This is indeed off-topic, but I do not think we have to hide anything - everyone knows that almost all of oVirt's development is done by Red Hat people. For many years, the oVirt project struggled with our scheduling and release cycles - we wanted to be able to use and rely on newer features of RHEL, when available, but also to do a decent job for oVirt users, who do not have access to RHEL the moment it's GA. As developers, and also as Red Hatters, we already had access to internal development versions of RHEL, so could QA on them prior to releases (of both RHEL and oVirt/RHV). But what to do with oVirt users? It was always hard. We often ended up rebuilding RHEL SRPMs right after their release, in copr, before CentOS included them, for oVirt - including adding these to ovirt-release* rpms, and including having to later remove them from there once CentOS did include them - just so that oVirt users can use them before they are available in CentOS Linux. With the move to CentOS Stream, all of this is gone. oVirt is only building upon Stream, always has the latest bits, and everyone is happy! Well, at least in theory. In practice, everything is still new to all of us, but I think that generally speaking, things are working well. To answer your direct question: RHV 4.4.6 is based on RHEL+AV 8.4, and the release dates of each part were coordinated internally for this. Best regards, -- Didi

On Mon, Jul 26, 2021 at 7:38 PM Gianluca Cecchi <gianluca.cecchi(a)gmail.com&gt; wrote: Thanks for the update! Best regards, -- Didi