Hide Forgot
Description of problem: Currently there is a failure in the validation flow, meaning, the all-in-one plugin is running (creating a dc and trying to create a cluster) even if cpu capabilities are missing from the machine.
ohad - is this with fedora 19? you need latest fedora qemu/libvirt for this to work. I don't think its worth handling this use case otherwise (unless i'm missing something)
actually it happened to me on a centos 6.4 machine.
how come there are not cpu capabilities?
because I accidently ran the setup on a vm obviously it can't run all in one but the setup should notify me and quit when this happens.
(In reply to Ohad Basan from comment #4) > because I accidently ran the setup on a vm > obviously it can't run all in one > but the setup should notify me and quit when this happens. I think that if nested VM is working (vdsm-hook-nestedvt), you should be able to install AIO also inside a VM (nested VM doesn't work on my laptop with emulated CPU westmere). It just seems that the all-in-one/cpu plugin doesn't detect correctly the VM's cpu.
Ohad Please reproduce and attach logs of the failure. Also, if you could change the code to pring that cpu value, that would be great too. Use something like self.logger.debug('The value of cpu is \'{cpu}\''.format(cpu=cpu)) Thanks.
Still missing the backport to 3.3 branch
Targeting to 3.3.2 sincce it's not in 3.3.1 branch.
rc build done, moving to ON_QA, packages will be available on updates-testing repository.
oVirt 3.3.2 has been released resolving the problem described in this bug report.