[pve-devel] copy_vm: new option -target

Alexandre DERUMIER aderumier at odiso.com
Fri May 3 10:56:37 CEST 2013


>>Linked clones are buggy with bobtail and will not be fixed before 
>>cuttlefish. So right now they're not usable. Also i'm mainly using 
>>templates with empty disks so they're not really linked at all i just 
>>use the templates as predefined settings. 

Oh ok, good to known !


As workaround, for now, maybe can you use qcow2 file as template ?
They use no space, and I think it should work to create sparse rbd target volume


----- Mail original ----- 

De: "Stefan Priebe - Profihost AG" <s.priebe at profihost.ag> 
À: "Alexandre DERUMIER" <aderumier at odiso.com> 
Cc: pve-devel at pve.proxmox.com, "Dietmar Maurer" <dietmar at proxmox.com> 
Envoyé: Vendredi 3 Mai 2013 10:14:08 
Objet: Re: [pve-devel] copy_vm: new option -target 

Hi, 

Am 03.05.2013 10:11, schrieb Alexandre DERUMIER: 
> Hi Stefan, 
> 
> Just to known, why don't you use linked clone ? (Do you have tried it, does it work fine with rbd ? are io/s good ?) 

Linked clones are buggy with bobtail and will not be fixed before 
cuttlefish. So right now they're not usable. Also i'm mainly using 
templates with empty disks so they're not really linked at all i just 
use the templates as predefined settings. 

> Another idea for later, we could also implement: 
> 1)rbd clone 
> 2)rbd flatten (this detach the clone from the parent, copying data from parent to child) 
> 
> 
> Same can be use with nexenta for "full clone" from a snapshot. 
> 
> 
> I could be also usefull to detach an existing linked clone (it's also possible with qcow2 clone I think) 
Yes that might be a really cool feature. 

Stefan 

> 
> ----- Mail original ----- 
> 
> De: "Dietmar Maurer" <dietmar at proxmox.com> 
> À: "Alexandre DERUMIER" <aderumier at odiso.com> 
> Cc: pve-devel at pve.proxmox.com, "Stefan Priebe - Profihost AG" <s.priebe at profihost.ag> 
> Envoyé: Vendredi 3 Mai 2013 08:42:25 
> Objet: RE: [pve-devel] copy_vm: new option -target 
> 
>>>> Should be possible to fix qemu-img? 
>> Yes, I think it should be possible. Code is not so huge. 
>> 
>> But Maybe do you have a better understanding of the code. (As you have played 
>> with backup ;) 
> 
> Maybe, but I need to make the other things stable first (this is only an 'optimization') 
> 



More information about the pve-devel mailing list