Bug 674535

Summary: virt-manager provides no way to set <driver type='qcow2'/>
Product: Red Hat Enterprise Linux 5 Reporter: Dan Yasny <dyasny>
Component: virt-managerAssignee: Cole Robinson <crobinso>
Status: CLOSED WONTFIX QA Contact: Virtualization Bugs <virt-bugs>
Severity: medium Docs Contact:
Priority: medium    
Version: 5.6CC: acathrow, berrange, dag, dallan, eblake, jwest, mjenner, mliu, mzhan, nzhang, rjones, rwu, virt-maint, xen-maint, zpeng
Target Milestone: rc   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-10-14 22:30:57 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
creating volume
none
attaching volume to VM
none
virt-manager.log none

Description Dan Yasny 2011-02-02 11:25:08 UTC
Description of problem:
After update to 5.6, all the qcow2 based VMs stopped booting. Manually starting them with the disk type set to qcow2 works, but updating the xml manually doesn't help, it gets overwritten to 'raw' by virt-manager and virsh, with no consideration for the actual disk type.

Created a new volume of the qcow2 type and attached to domain - still can see it set to 'raw' in the xml for that domain

Tested same on F14 - seems to be working correctly

Version-Release number of selected component (if applicable):
RHEL5.6
libvirt-0.8.2-15.el5_6.1
libvirt-devel-0.8.2-15.el5_6.1
libvirt-devel-0.8.2-15.el5_6.1
libvirt-python-0.8.2-15.el5_6.1
libvirt-0.8.2-15.el5_6.1


How reproducible:
Always on 5.6, used to work on 5.5

Steps to Reproduce:
1.create a new qcow2 volume
2.create a new VM, attach the qcow2 vol.
3.start VM and check the xml file - 
  
Actual results:
will see the VM not booting and xml has 'raw' in the disk settings

Expected results:
volume settings should not be overwritten to 'raw' in the xml file

Additional info:
reproduced internally.

Comment 1 Dan Yasny 2011-02-02 11:33:33 UTC
Also tried to edit the xml manually, change type back to qcow2. It gets overwritten back to 'raw' anyway

Comment 2 Daniel Berrangé 2011-02-02 11:45:24 UTC
I'm not able to reproduce this problem. I tried the following sequence:

1. virsh start rhel5clone

2. cat > disk.xml <<EOF
    <disk type='file' device='disk'>
      <driver name='qemu' type='qcow2'/>
      <source file='/var/lib/libvirt/images/extra1.img'/>
      <target dev='vda' bus='virtio'/>
    </disk>
EOF

3. virsh attach-device rhel5clone disk.xml

4. virsh dumpxml rhel5clone

<domain type='kvm' id='1'>
  <name>rhel5clone</name>
  <uuid>8c286c95-4828-fa98-e84d-81f034025241</uuid>
  <memory>419200</memory>
  <currentMemory>418816</currentMemory>
  <vcpu>1</vcpu>
  <os>
    <type arch='x86_64' machine='rhel5.4.0'>hvm</type>
    <boot dev='hd'/>
  </os>
  <features>
    <acpi/>
    <apic/>
    <pae/>
  </features>
  <clock offset='utc'>
    <timer name='pit' tickpolicy='delay'/>
  </clock>
  <on_poweroff>destroy</on_poweroff>
  <on_reboot>restart</on_reboot>
  <on_crash>restart</on_crash>
  <devices>
    <emulator>/usr/libexec/qemu-kvm</emulator>
    <disk type='file' device='disk'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/rhel5clone.img'/>
      <target dev='hda' bus='ide'/>
      <alias name='ide0-hd0'/>
      <address type='drive' controller='0' bus='0' unit='0'/>
    </disk>
    <disk type='file' device='disk'>
      <driver name='qemu' type='qcow2'/>
      <source file='/var/lib/libvirt/images/extra1.img'/>
      <target dev='vda' bus='virtio'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x2b5a'/>
    </disk>
    <controller type='ide' index='0'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/>
    </controller>
    <interface type='bridge'>
      <mac address='00:16:36:2d:45:a9'/>
      <source bridge='xenbr1'/>
      <ip address='192.168.122.5'/>
      <target dev='vnet0'/>
      <filterref filter='clean-traffic'>
        <parameter name='IP' value='192.168.122.5'/>
      </filterref>
    </interface>
    <serial type='pty'>
      <source path='/dev/pts/2'/>
      <target port='0'/>
    </serial>
    <console type='pty' tty='/dev/pts/2'>
      <source path='/dev/pts/2'/>
      <target port='0'/>
    </console>
    <video>
      <model type='cirrus' vram='9216' heads='1'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/>
    </video>
  </devices>
</domain>


It still reports qcow2 for the 2nd disk as expected.

In addition using 'define' works too

1. virsh dumpxml rhel5clone > a.xml
2. virsh destroy rhel5clone
3. virsh define a.xml
4. virsh dumpxml rhel5clone

still reports 'qcow2'.

What is the precise sequence of virsh operations used to reproduce the problem ?

Comment 3 Dan Yasny 2011-02-02 12:34:53 UTC
editing the xml with virsh edit works

attaching the screenshots of my steps virt1 and virt2.png, their result is:
    <disk type='file' device='disk'>
      <driver name='qemu' type='raw'/>
      <source file='/var/lib/libvirt/images/test123.img'/>
      <target dev='hda' bus='ide'/>
      <address type='drive' controller='0' bus='0' unit='0'/>
    </disk>

Comment 4 Dan Yasny 2011-02-02 12:35:44 UTC
Created attachment 476551 [details]
creating volume

Comment 5 Dan Yasny 2011-02-02 12:36:29 UTC
Created attachment 476552 [details]
attaching volume to VM

Comment 6 Daniel Berrangé 2011-02-02 12:46:35 UTC
Ok, this is likely a virt-manager problem, because that sequence you show probably relies on QEMU being able to probe the volume format. Please attach the $HOME/.virt-manager/virt-manager.log file.

Comment 7 Dan Yasny 2011-02-02 13:04:25 UTC
http:/10.35.24.10/share/virt-manager.log
I've cut it down to only todays' events

Comment 8 Richard W.M. Jones 2011-02-02 13:19:05 UTC
(In reply to comment #7)
> http:/10.35.24.10/share/virt-manager.log
> I've cut it down to only todays' events

I can't get to that (even after correcting the obvious
mistake in the URL).  Can you attach it instead?

Comment 9 Dan Yasny 2011-02-02 13:37:23 UTC
Created attachment 476568 [details]
virt-manager.log

Comment 12 Cole Robinson 2011-03-10 15:50:54 UTC
*** Bug 643234 has been marked as a duplicate of this bug. ***

Comment 14 Cole Robinson 2011-06-15 14:43:27 UTC
Yeah this slipped of my radar, deferring to 5.8

Comment 16 Cole Robinson 2011-10-14 22:30:57 UTC
After evaluating this, providing UI to change this value for existing disk images would be way to invasive for the old old version of virt-manager in RHEL5. Also given that there hasn't been any customer complaints about this, I don't think this is worth backporting for RHEL5, the only real option is a rebase which likely isn't going to happen either. This is all properly fixed in RHEL6.

Any users that are having issues will need to use 'virsh edit' to fix things.

Closing as WONTFIX for RHEL5