Bug 627143

Libvirt haven't check image real format instead of passing default raw format
to qemu-kvm, the following information is come from libvirtd.log:

......
16:27:37.471: debug : qemuMonitorAddDrive:1897 : mon=0x1ef3890
drive=file=/var/lib/libvirt/images/mytest.qcow2,if=none,id=drive-virtio-disk1,format=raw
......
16:27:37.471: debug : virJSONValueToString:1059 :
result={"execute":"__com.redhat_drive_add","arguments":{"file":"/var/lib/libvirt/images/mytest.qcow2","id":"drive-virtio-disk1","format":"raw"}}
16:27:37.471: debug : qemuMonitorJSONCommandWithFd:217 : Send command
'{"execute":"__com.redhat_drive_add","arguments":{"file":"/var/lib/libvirt/images/mytest.qcow2","id":"drive-virtio-disk1","format":"raw"}}'

The error report: unsupported driver name 'phy', maybe it's the focal point.

set driver name='qemu', keep type=raw
# virsh edit demo
...
<driver name='qemu' type='raw'/>
...

# virsh start demo
Domain demo started

But when I only change 'raw' to 'qcow2', guest still start fail.

# virsh edit demo
...
<driver name='phy' type='qcow2'/>
...

# virsh start demo
error: Failed to start domain demo
error: internal error unsupported driver name 'phy' for disk '/var/lib/libvirt/images/foo.img'

(In reply to comment #2)
> Libvirt haven't check image real format instead of passing default raw format
> to qemu-kvm, the following information is come from libvirtd.log:
> 
> ......
> 16:27:37.471: debug : qemuMonitorAddDrive:1897 : mon=0x1ef3890
> drive=file=/var/lib/libvirt/images/mytest.qcow2,if=none,id=drive-virtio-disk1,format=raw
> ......
> 16:27:37.471: debug : virJSONValueToString:1059 :
> result={"execute":"__com.redhat_drive_add","arguments":{"file":"/var/lib/libvirt/images/mytest.qcow2","id":"drive-virtio-disk1","format":"raw"}}
> 16:27:37.471: debug : qemuMonitorJSONCommandWithFd:217 : Send command
> '{"execute":"__com.redhat_drive_add","arguments":{"file":"/var/lib/libvirt/images/mytest.qcow2","id":"drive-virtio-disk1","format":"raw"}}'

we can use 'subdriver=qcow2' in command attach-device, but it's not the workaround of the bug.

# virsh attach-disk demo /var/lib/libvirt/images/foo.img vdb --subdriver qcow2
Disk attached successfully

# virsh dumpxml demo
...
<disk type='block' device='disk'>
      <driver name='phy' type='qcow2'/>
      <source dev='/var/lib/libvirt/images/foo.img'/>
      <target dev='vdb' bus='virtio'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/>
</disk>

Find some mismatch between attach-disk and xml.

Juxtapose the candidate of --driver in 'man virsh' and the actual value in xml, 
I think the value of --driver should be matched <disk type='phy' device='disk'>,
but it pass the value to <driver name='phy' type='qcow2'>.

# man virsh
...
    attach-disk domain-id source target optional --driver driver --subdriver subdriver --type type --mode mode

Attach a new disk device to the domain.  source and target are paths for the files and devices.  
>>>>
 driver can be file, tap or phy depending on the kind of access.
<<<  
>>>>
 type can indicate cdrom or floppy as alternative to the disk default
<<<,
although this use only replaces the media within the existing virtual cdrom or floppy device; consider using update-device for this usage instead.  mode can specify the two specific mode readonly or shareable.

# virsh edit demo
...
<disk type='file' device='disk'>
      <driver name='qemu' type='qcow2' cache='none'/>
      <source file='/var/lib/libvirt/images/RHEL-Server-6-64-virtio.qcow2'/>
      <target dev='hda' bus='ide'/>
      <address type='drive' controller='0' bus='0' unit='0'/>
</disk>
<disk type='file' device='cdrom'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/cdrom1.iso'/>
      <target dev='hdc' bus='ide'/>
      <readonly/>
      <address type='drive' controller='0' bus='1' unit='0'/>
</disk>

I test it on rc4 to attach a raw disk, it also has problem and cat not start
#virsh attach-disk demo /var/lib/libvirt/images/foo.img vdb
#virsh dumpxml demo
...
    <disk type='block' device='disk'>
      <driver name='phy' type='raw'/>
      <source dev='/var/lib/libvirt/images/foo.img'/>
      <target dev='vdb' bus='virtio'/>
      <alias name='virtio-disk1'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
    </disk>
...

FWIW I have the same problem on Fedora 14 beta.  I have added a /dev/md0 device to my guest and it works fine until I shut down.  To add insult to the injury, I can't detach the disk from the virsh console using 'detach-disk' because the domain isn't running.  But I can't start the domain.  Luckily, I can remove the device from virt-manager.

Is there a Fedora 14 bug I should subscribe to in order to follow the progress?

missed one upstream patch which is to address this problem: http://www.redhat.com/archives/libvir-list/2010-September/msg00156.html

though it's not addressed completely, depends on user must use the new options introduced by this patch.

- Osier

patch pushed upstream:

http://www.redhat.com/archives/libvir-list/2010-December/msg00097.html

- Osier

Hi Osier,

Should the patch be backported into RHEL5.6/5.7 for Xen? Xen could handle the situation when no driver was given as a default(see BZ#658712), so it is ok for libvirt not to give a default driver value.

Yufang

Built into libvirt-0.8.7-1.el6

verify PASS on libvirt-0.8.7-1.el6

after start a guest do
# virsh attach-disk test /var/lib/libvirt/images/foo.img vdb
Disk attached successfully

# virsh dumpxml test
...
<disk type='block' device='disk'>
      <driver name='qemu' type='raw'/>
      <source dev='/var/lib/libvirt/images/foo.img'/>
      <target dev='vdb' bus='virtio'/>
      <alias name='virtio-disk1'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
</disk>
...

After shutdown, the device "foo.img" is removed.

# virsh start test
Domain test started

no error occured. 

qemu-kvm-0.12.1.2-2.128.el6.x86_64
kernel-2.6.32-94.el6.x86_64
libvirt-0.8.7-1.el6

*** Bug 632580 has been marked as a duplicate of this bug. ***

Blocked by new bug 676548 and will verify it later.

Rechecked it on the following test environment according to comment 13. It is passed.

Test environment:
libvirt-0.8.7-6.el6
qemu-kvm-0.12.1.2-2.144.el6
kernel-2.6.32-113.el6

tested with 
libvirt-0.8.7-18.el6.x86_64
qemu-kvm-0.12.1.2-2.158.el6.x86_64
kernel-2.6.32-131.0.1.el6.x86_64

after hot-plug disk , shutdown guest works fine , and guest can be start after stop . 

So keep bug status as VERIFIED

    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
virsh set a "phy" as the disk driver if user doesn't specify one when attaching a disk (virsh attach-disk), this breaks the domain start after attaching a disk without specifying the driver, this update fixed the problem.

    Technical note updated. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    Diffed Contents:
@@ -1 +1 @@
-virsh set a "phy" as the disk driver if user doesn't specify one when attaching a disk (virsh attach-disk), this breaks the domain start after attaching a disk without specifying the driver, this update fixed the problem.+virsh sets "phy" as the disk driver if user doesn't specify one when attaching a disk (virsh attach-disk), this breaks the domain start after attaching a disk without specifying the driver, this update fixed the problem.

    Technical note updated. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    Diffed Contents:
@@ -1 +1 @@
-virsh sets "phy" as the disk driver if user doesn't specify one when attaching a disk (virsh attach-disk), this breaks the domain start after attaching a disk without specifying the driver, this update fixed the problem.+If the user did not specify a disk driver when hot-plugging a disk with "virsh attach-disk", virsh set "phy" as the driver value by default. Because this value is not supported everywhere, the disk did not persist over domain shutdown, and could prevent domain startup. This update corrects virsh behavior such that the driver value is not set if it is not provided by the user.

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHBA-2011-0596.html