(In reply to wanghui from comment #0)
> Description of problem:
> Self engine vm hangs on "starting rhsmcerted" process When rhevh choose
> install vm from OVA file.
> 
> Version-Release number of selected component (if applicable):
> rhev-hypervisor7-7.0-20141006.0.el7ev
> ovirt-node-3.1.0-0.20.20141006gitc421e04.el7.noarch.rpm
> ovirt-node-plugin-hosted-engine-0.2.0-2.0.el7.x86_64
> ovirt-hosted-engine-setup-1.2.1-1.el7.noarch
> ovirt-host-deploy-1.3.0-1.el7.noarch
> ovirt-hosted-engine-ha-1.2.2-2.el7.noarch
> rhevm-appliance-20141006.0-1.x86_64.rhevm.ova
> 
> How reproducible:
> 100%
> 
> Steps to Reproduce:
> 1. Install rhev-hypervisor7-7.0-20141006.0.el7ev.
> 2. Configure hosted engine using OVA option.
> 3. VNC connect engine vm.
> 
> Actual results:
> 1. After step3, the self engine vm hangs on "starting rhsmcerted" proces.
> 
> Expected results:
> 1. VM should be started.
> 
> Additional info:

Hello wanghui,

Could you please provide the logs from the system?

/var/log/ovirt*
/var/log/ovirt-hosted-engine*
/var/log/message

Thanks!

Created attachment 947413 [details]
log files

Created attachment 947425 [details]
screen shot for ova vm

Sandro, can you tell how serious this is?

As far as I know this bug has not popped up in the latest testing, so please re-test this with todays iso and put it back to ASSIGNED if necessary.

I don't think this related at all with hosted engine.

I see in vdsm logs:

Thread-102::DEBUG::2014-10-16 05:08:09,423::libvirtconnection::143::root::(wrapper) Unknown libvirterror: ecode: 74 edom: 10 level: 2 message: argument unsupported: QEMU driver does not support <metadata> element
Thread-102::ERROR::2014-10-16 05:08:09,424::vm::373::vm.Vm::(_sampleCpuTune) vmId=`a1f1529d-5c27-42a6-bb3f-bad524b792e2`::libvirt does not support metadata

which seems to be something new and I think it requires an additional BZ to be tracked.

After the above error, the VM is still working, I see a graphic event later when the console connects to qemu:

libvirtEventLoop::DEBUG::2014-10-16 05:08:26,764::clientIF::551::vm.Vm::(dispatchLibvirtEvents) vmId=`a1f1529d-5c27-42a6-bb3f-bad524b792e2`::graphics event phase 0 localAddr {'node': '0.0.0.0', 'service': '5900', 'family': 0} remoteAddr {'node': '10.66.65.29', 'service': '49726', 'family': 0}authScheme vnc subject []


As far as I can tell, the VM is still alive and it's probably stuck due to something within the OVA or just waiting on resources from the host (high load / low memory / swapping / whatever).
I suggest to check the VM logs (access the VM disk from somewhere else or boot the VM from iso).

(In reply to Sandro Bonazzola from comment #6)
> I don't think this related at all with hosted engine.
> 
> I see in vdsm logs:
> 
> Thread-102::DEBUG::2014-10-16
> 05:08:09,423::libvirtconnection::143::root::(wrapper) Unknown libvirterror:
> ecode: 74 edom: 10 level: 2 message: argument unsupported: QEMU driver does
> not support <metadata> element
> Thread-102::ERROR::2014-10-16 05:08:09,424::vm::373::vm.Vm::(_sampleCpuTune)
> vmId=`a1f1529d-5c27-42a6-bb3f-bad524b792e2`::libvirt does not support
> metadata
> 
> which seems to be something new and I think it requires an additional BZ to
> be tracked.

This is about an older version of libvirt and mom trying to adjust cpu usage.
This should be tested with updated libvirt, mom and vdsm of 3.5.0.

verified on rhev-hypervisor7-7.0-20141202.0 build, this kind of hang issue does not appear now.

ovirt-node-plugin-hosted-engine-0.2.0-5.0.el7ev.x86_64
ovirt-hosted-engine-setup-1.2.1-6.el7ev.noarch
ovirt-hosted-engine-ha-1.2.4-2.el7ev.noarch
ovirt-node-3.1.0-0.28.20141126git25ce016.el7.noarch
rhevm-appliance-20141112.0-1.x86_64.rhevm.ova

RHEV 3.5.0 has been released. I am closing this bug, because it has been VERIFIED.