Description of problem: we need to allow over-commit the cpu number of a VM over a host. this is supported and doesn't pose a technical issue. Version-Release number of selected component (if applicable): How reproducible: 100% Steps to Reproduce: 1.Edit a running Vm 2. the numnber of sockets will not exceed the overall number of the host Actual results: the actual max numnber of CPU's should be the the vdc_config value, which is the qemu technical spec max http://doc.opensuse.org/products/draft/SLES/SLES-kvm_sd_draft/cha.kvm.limits.html Expected results: the limitations should be config based Additional info:
roy - why would we allow over-commit in hot-plug, when we wouldn't allow it when running the VM to begin with. also, in such a bug its important to be accurate: 1. why sockets and not cores? 2. cores or hyperthreads based on cluster policy?
we should allow it based on policy (the default one disallows, but that can be changed). We currently do not consult the scheduler at all. indeed it's about vCPUs rather than socket/core/thread. Those should only comply with the maximum topology limits(16/16/?)
seems we're out of time on 3.4.z really, so let's try to finish this up at least for 3.5
Moving pending bugs not fixed in 3.5.0 to 3.5.1.
This is an automated message: This bug should be fixed in oVirt 3.5.1 RC1, moving to QA
oVirt 3.5.1 has been released. If problems still persist, please make note of it in this bug report.