On 06/16/2014 02:06 AM, Tuân Tạ Bá wrote:
> Hi everyone,
>
> I use KVM on Openstack Cloud. And KVM migrated successfull, but not ping.
>

> *Source Host SIDE:  *"/var/log/libvirt/libvirtd.log"

>
> 2014-06-16 05:15:50.599+0000: 2096: warning :
> qemuMigrationCancelDriveMirror:1383 : Unable to stop block job on
> drive-virtio-disk0

> 2014-06-16 05:15:50.601+0000: 2096: *warning :

> qemuMigrationCancelDriveMirror:1383 : Unable to stop block job on

> drive-virtio-disk1*

These warnings don't look good; have you investigated why they are
occurring?

> *2014-06-16 05:15:50.651+0000: 2096: error : virCommandWait:2348 : internal

> error: Child process (ovs-vsctl --timeout=5 get Interface tap6977e4a1-cd
> external_ids:PortData) unexpected exit status 1: ovs-vsctl: no key

> "PortData" in Interface record "tap6977e4a1-cd" column external_ids*

>
> 2014-06-16 05:15:50.651+0000: 2096: error :
> virNetDevOpenvswitchGetMigrateData:219 : Unable to run command to get OVS
> port data for interface tap6977e4a1-cd: Operation not permitted
> 2014-06-16 05:15:50.651+0000: 2096: error :
> qemuMigrationCookieNetworkAlloc:367 : Unable to run command to get OVS port
> data for interface tap6977e4a1-cd: Operation not permitted

Based on this failure, migration should not be successful.

> 2014-06-16 05:15:50.651+0000: 2096: warning : qemuMigrationRun:3262 :
> Unable to encode migration cookie
> 2014-06-16 05:15:53.321+0000: 2096: warning : virAuditSend:135 : Failed to
> send audit message virt=kvm op=stop reason=migrated vm="instance-0001b37c"
> uuid=631ac4c7-8e84-4cf6-ad98-93877cdbe1bd vm-pid=-1: Operation not permitted

That doesn't look good either.  Are you sure your system is set up
correctly to allow audit messages to be generated?  But it's probably
unrelated to the failure you are seeing.

>
> *Destination Host SIDE:* "/var/log/libvirt/libvirtd.log"

> ......
>
> 2014-06-16 05:15:41.419+0000: 2229: warning : virAuditSend:135 : Failed to
> send audit message virt=kvm resrc=cgroup reason=allow vm="instance-000
> 1b37c" uuid=631ac4c7-8e84-4cf6-ad98-93877cdbe1bd
> cgroup="/sys/fs/cgroup/devices/machine/instance-0001b37c.libvirt-qemu/"
> class=major category=pty
>  maj=88 acl=rw: Operation not permitted
> 2014-06-16 05:15:41.620+0000: 2229: warning :
> qemuDomainObjEnterMonitorInternal:1175 : This thread seems to be the async
> job owner; entering moni
> tor without asking for a nested job is dangerous
> 2014-06-16 05:15:43.552+0000: 2229: warning : virAuditSend:135 : Failed to
> send audit message virt=kvm resrc=vcpu reason=start vm="instance-0001b
> 37c" uuid=631ac4c7-8e84-4cf6-ad98-93877cdbe1bd old-vcpu=0 new-vcpu=2:
> Operation not permitted

> 2014-06-16 05:15:43.552+0000: 2229: warning : virAuditSend:135 :* Failed to
> send audit message *virt=kvm op=start reason=migrated vm="instance-0001
> b37c" uuid=631ac4c7-8e84-4cf6-ad98-93877cdbe1bd vm-pid=21063:* Operation
> not permitted*

>
> 1) I'm using libvirt version 1.1.1,
>
> libvirt-bin      1.1.1-0ubuntu8.5~cloud0
> libvirt0      1.1.1-0ubuntu8.5~cloud0
> python-libvirt      1.1.1-0ubuntu8.5~cloud0
>
> 2) #cat /etc/libvirt/qemu.conf
> ...
> user = "root"
> group = "root"
>
> I searched GOOGLE, René speaks "downgrading to libvirt 1.1.0 and
> reinstalling everything. They will ok". But I don't think that.
>
> How to resolve this error?.

We've had some bug fixes for migration in the meantime; can you
reproduce the error with libvirt 1.2.5?

--
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org