Roman Mohr
Roman Mohr
> @rmohr > Currently, the mapping between a VM SR-IOV interface and the underlying VFs is not consistent when more than one VF is from the same resource pool. >...
> Hey @rmohr, thanks to @ormergi explained the issue in short. Let me give you some more detail that may explain this change. If you have more questions, there is...
> @ormergi correct me if I'm wrong, but I think the main issue with doing it on vmi-status is that we may get in trouble on migration. AFAIU, when migrating...
> I would prefer not to see the information in the VMI status be used as a desired state input. Per my understanding of the pattern, spec is where the...
> > Thanks > > Since it is blocking kubevirtci and cause kubevirt nightly conformance to fail, what about meanwhile unblocking those by adding a check if the nodes are...
Yes we have the side-effect already but not when using sonobuoy. Two bugs cancelling out each other. 😁
@xpivarc @maiqueb @oshoval will ping you on slack. let's discuss this quickly.
> Currently, there are three place to do CPU arch specific modification, > > 1. in the `virt-api webhooks`, set DefaultCPUModel, DefaultBootloader and DefaultDisksBus > 2. in the `virt-config/configuration.go`, set...