Rephrasing mail send by Andrea Bolognani to internal mailing list: It looks like KubeVirt is currently[1] installing libvirt using a very straightforward dnf invocation: ENV LIBVIRT_VERSION 5.1.0 RUN dnf install -y dnf-plugins-core && \ dnf copr enable -y @virtmaint-sig/virt-preview && \ dnf install -y \ libvirt-daemon-kvm-${LIBVIRT_VERSION} \ libvirt-client-${LIBVIRT_VERSION} \ ... This will drag in a bunch of libvirt components which might not be useful to KubeVirt, such as for example nwfilter and device assignment support, as well as drivers for a number of storage technologies that IIUC are not going to be used. Suggestion is to look into tailoring the above to KubeVirt's actual requirements, which would probably result in something more along the lines of dnf install -y libvirt-daemon \ libvirt-daemon-driver-qemu \ libvirt-daemon-driver-secret \ libvirt-daemon-driver-storage-disk \ libvirt-client \ qemu-kvm and reduced disk and memory footprints, as well as attach surface. [1] https://github.com/kubevirt/libvirt/blob/master/Dockerfile
I've posted a PR implementing this change upstream: https://github.com/kubevirt/libvirt/pull/37 It builds on two other PRs that significantly change how the kubevirt/libvirt container image is generated, and which have not been merged yet, so it could take a while before the change makes its way upstream; it also needs significantly more testing than what I have been able to give it on my own. Additionally, I'm unclear on what needs to happen, once the change has been approved upstream, before it can benefit Red Hat customers: I assume it will have to be backported somehow... Anyway, things are moving :)
This has now been merged upstream: https://github.com/kubevirt/libvirt/commit/d6cd8894a5c13ddb8ca03379860238f9043eec22 I also opened a PR to switch KubeVirt itself to use the new image: https://github.com/kubevirt/kubevirt/pull/2890
Andrea, many thanks for contributing this patch. Petr, please work with Andrea through the process to make sure that kubevirt master will see this enhancement.
(In reply to Fabian Deutsch from comment #4) > Petr, please work with Andrea through the process to make sure that kubevirt > master will see this enhancement. It's in master already: https://github.com/kubevirt/kubevirt/commit/a174131cea8ae3362b58b9b6f4f48992efa21f8d What I'm unclear about is the downstream process, but I assume you guys have that part figured out already and don't necessarily need me to be involved in it... I'm always happy to help where I can, though :)
I see Andreas PR merged, thank you for that! I might have a question or two for downstream though. Nevertheless, good one Andrea.
The downstream PR has been merged. I believe this bug can go to POST.
Verify with: oc get kv -n openshift-cnv -o yaml | grep -i operatorversion operatorVersion: v0.26.1 Create vm and check libvirt at compute pod. [root@vm-cirros /]# rpm -qa | grep libvirt libvirt-libs-5.6.0-10.module+el8.1.1+5309+6d656f05.x86_64 libvirt-daemon-5.6.0-10.module+el8.1.1+5309+6d656f05.x86_64 libvirt-client-5.6.0-10.module+el8.1.1+5309+6d656f05.x86_64 libvirt-daemon-driver-qemu-5.6.0-10.module+el8.1.1+5309+6d656f05.x86_64 libvirt-bash-completion-5.6.0-10.module+el8.1.1+5309+6d656f05.x86_64
Moving this BZ back to NEW and deferring to the next release. We removed too many components, which caused other issues: https://bugzilla.redhat.com/show_bug.cgi?id=1810493
Test runs with new Kernel RHEL 8.2 are passing Migration is working. Moving to verify.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2020:3194