[Gluster-users] IP failover in QEMU-Glusterfs native integration using libgfapi?

[Chenggang ZHANG]

Hi Dan,

Even with Glusterfs 4.3.2, we still need some extra steps to make native 
integration with QEMU work, you can find more details from 
http://www.gluster.org/community/documentation/index.php/Building_QEMU_with_gfapi_for_Debian_based_systems

As for the qcow2 image importing, you can do it in two ways:
(1) The FUSE approach:
       mount -t glusterfs //brickServerIp/volumeName /mnt
       qemu-img convert -f raw -O qcow2 
path-to-the-source-raw-image-file /mnt/imageName.qcow2
(2) The native approach, e.g., qemu-img convert -f raw -O qcow2 
path-to-the-source-raw-image-file 
gluster://brickServerIp/volumeName/imageName.qcow2

Good luck!

Best Regards,
Bob

On 04/02/2014 12:26 PM, Daniel wrote:
> Hi,
>
> I am trying to achieve the same thing.   Why did you have to enable 
> QEMU native integration when Glusterfs 4.3.2 already has native QEMU 
> support?
>
> How did you import the qcow2 image and how did you set it up.
>
>
> Thanks,
>
> Dan
>
> On 04/02/2014 10:59 AM, Chenggang ZHANG wrote:
>> Dear All,
>>
>> I am using QEMU 1.7.0 and GlusterFs 4.3.2 and have carried out 
>> necessary configurations to enable QEMU-Glusterfs native integration 
>> (i.e., using libgfapi instead of FUSE).
>>
>> On the server side, I have configured a volume named "image" using 
>> two bricks on two different machines "brk1" and "brk2". And a qcow2 
>> image *centos.qcow2* has been imported to Glusterfs.
>>
>> On the client side (e.g., the KVM host), the configuration "-drive 
>> file=gluster://*brk1*/image/centos.qcow2" is used when starting a KVM 
>> VM. The VM boots successfully.
>> When it's running, I shut down the machine brk1 on purpose and want 
>> to see whether the KVM VM can automatically switch to use "-drive 
>> file=gluster://*brk2*/image/centos.qcow2" (i.e., IP failover). 
>> Unfortunately, the VM crashes.
>>
>> I am writing to ask whether the feature "IP failover" in 
>> QEMU-Glusterfs native integration is supported. Thanks so much for 
>> your kind help!
>>
>>
>> Best Regards,
>> Bob ZHANG
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://supercolony.gluster.org/mailman/listinfo/gluster-users
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140402/6cd04d0f/attachment.html>