Bug 785683

Summary: A live snapshot shouldn't reconfigure the backing file path in the new image
Product: Red Hat Enterprise Linux 6 Reporter: Federico Simoncelli <fsimonce>
Component: qemu-kvmAssignee: Federico Simoncelli <fsimonce>
Status: CLOSED ERRATA QA Contact: Virtualization Bugs <virt-bugs>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 6.3CC: abaron, acathrow, bsarathy, danken, ehabkost, juzhang, kwolf, mkenneth, pbonzini, shu, tburke, virt-maint
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: qemu-kvm-0.12.1.2-2.264.el6 Doc Type: Bug Fix
Doc Text:
No documentation needed
 Story Points: ---
Clone Of: Environment:
Last Closed: 2012-06-20 08:52:57 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On:    
Bug Blocks: 806280    

Description Federico Simoncelli 2012-01-30 10:49:59 UTC 
Description of problem:
During a live snapshot (snapshot_blkdev) if the target qcow image already exists it is re-initialized by qemu-kvm overriding any previous configuration.
This is a problem for VDSM that is preparing the VM images on the SPM configuring the backing file path (where needed) to a relative path.

Version-Release number of selected component (if applicable):
qemu-kvm-0.12.1.2-2.221.el6.x86_64

How reproducible:
100%

Steps to Reproduce:

# qemu-img create -f qcow2 /tmp/vmdisk1.qcow2 4G
Formatting '/tmp/vmdisk1.qcow2', fmt=qcow2 size=4294967296 encryption=off cluster_size=65536 

# qemu-img create -f qcow2 -b vmdisk1.qcow2 /tmp/vmdisk1-snap1.qcow2 4G
Formatting '/tmp/vmdisk1-snap1.qcow2', fmt=qcow2 size=4294967296 backing_file='vmdisk1.qcow2' encryption=off cluster_size=65536 

# qemu-img info /tmp/vmdisk1-snap1.qcow2 | grep backing
backing file: vmdisk1.qcow2 (actual path: /tmp/vmdisk1.qcow2)

# /usr/libexec/qemu-kvm -hda /tmp/vmdisk1.qcow2 -monitor stdio -vnc :5900
QEMU 0.12.1 monitor - type 'help' for more information
(qemu) snapshot_blkdev ide0-hd0 /tmp/vmdisk1-snap1.qcow2
Formatting '/tmp/vmdisk1-snap1.qcow2', fmt=qcow2 size=4294967296 backing_file='/tmp/vmdisk1.qcow2' backing_fmt='qcow2' encryption=off cluster_size=65536 
(qemu) quit
  
Actual results:
The backing file is an absolute path:
# qemu-img info /tmp/vmdisk1-snap1.qcow2 | grep backing
backing file: /tmp/vmdisk1.qcow2 (actual path: /tmp/vmdisk1.qcow2)

Expected results:
The backing file must be the relative path configured at the beginning:
# qemu-img info /tmp/vmdisk1-snap1.qcow2 | grep backing
backing file: vmdisk1.qcow2 (actual path: /tmp/vmdisk1.qcow2)

Additional info:
An initial patch for this issue was proposed in:

http://lists.gnu.org/archive/html/qemu-devel/2011-10/msg00234.html

The patch was rejected since it was completely avoiding the re-initialization of the new image.
A good compromise would be to re-initialize the new image re-using the same backing file path (when the image is already present and the path is correct).

Pseudocode:

  def snapshot_blkdev(device, new_image_file):
    img_backfile = get_device_file(device)

    if image_exists(new_image_file):
      tmp_backfile = get_image_backing_file(new_image_file)

      if same_path(img_backfile, tmp_backfile):
        img_backfile = tmp_backfile
      else:
        fail(wrong_backing_file)

    [...]
Comment 2 Dor Laor 2012-03-13 10:06:57 UTC 
According to Paolo libvirt works around it.
Federico, can we close the issue?
Comment 9 Shaolong Hu 2012-03-29 06:37:08 UTC 
Test on 267rhev:
-------------------

1. sn1 does not exist w/ absolute-paths:

{ "execute": "blockdev-snapshot-sync", "arguments": { "device": "drive-virtio-disk0", "snapshot-file": "/root/sn1", "format": "qcow2", "mode": "absolute-paths" } }

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 3.4M
cluster_size: 65536
backing file: /root/RHEL-Server-6.3-64-virtio.qcow2 (actual path: /root/RHEL-Server-6.3-64-virtio.qcow2)

2. sn1 exists w/ absolute-paths:

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 136K
cluster_size: 65536
backing file: base (actual path: base)

{ "execute": "blockdev-snapshot-sync", "arguments": { "device": "drive-virtio-disk0", "snapshot-file": "/root/sn1", "format": "qcow2", "mode": "absolute-paths" } }

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 1.5M
cluster_size: 65536
backing file: /root/RHEL-Server-6.3-64-virtio.qcow2 (actual path: /root/RHEL-Server-6.3-64-virtio.qcow2)

3. sn1 exists w/ existing:

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 136K
cluster_size: 65536
backing file: base (actual path: base)

{ "execute": "blockdev-snapshot-sync", "arguments": { "device": "drive-virtio-disk0", "snapshot-file": "/root/sn1", "format": "qcow2", "mode": "existing" } }

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 136K
cluster_size: 65536
backing file: base (actual path: base)

4. sn1 exists w/ no-backing-file:

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 136K
cluster_size: 65536
backing file: base (actual path: base)

{ "execute": "blockdev-snapshot-sync", "arguments": { "device": "drive-virtio-disk0", "snapshot-file": "/root/sn1", "format": "qcow2", "mode": "no-backing-file" } }

[root@shu ~]# qemu-img info sn1
image: sn1
file format: qcow2
virtual size: 20G (21474836480 bytes)
disk size: 2.0M
cluster_size: 65536


Conclusion:
------------
1.QMP works fine.
2.HMP has no "no-backing-file" option, i assume the [-n] is the equivalence to existing/absolute-paths:

snapshot_blkdev [-n] device [new-image-file] [format] -- initiates a live snapshot
			of device. If a new image file is specified, the
			new image file will become the new root image.
			If format is specified, the snapshot file will
			be created in that format. Otherwise the
			snapshot will be internal! (currently unsupported).
			The default format is qcow2.  The -n flag requests QEMU
			to reuse the image found in new-image-file, instead of
			recreating it from scratch.

3.above tests base on my understand of the patch, if no more test is needed and HMP issue is not a problem, i am willing to verify this bug.
Comment 10 Paolo Bonzini 2012-03-29 08:16:08 UTC 
no-backing-file is used only with mirroring, so snapshot_blkdev does not have it.  Testing looks good, thanks!
Comment 12 Dor Laor 2012-04-22 11:30:35 UTC 
    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:
No documentation needed
Comment 14 errata-xmlrpc 2012-06-20 08:52:57 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHBA-2012-1017.html