[pve-devel] vm migration + storage migration with ndb workflow notes

[Alexandre DERUMIER]

about qemu-img vs drive-mirror, 

I see a big speed difference

for a 32gb file, migrate on same local, qcow2 -> raw, both writeback

qemu-img : 30s
drive-mirror (with paused vm) : 5min


So I think drive-mirror code is more complex than qemu-img


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

De: "Alexandre DERUMIER" <aderumier at odiso.com> 
À: "Dietmar Maurer" <dietmar at proxmox.com> 
Cc: pve-devel at pve.proxmox.com 
Envoyé: Lundi 21 Janvier 2013 09:11:01 
Objet: Re: [pve-devel] vm migration + storage migration with ndb workflow notes 

>>Are you sure? Usually you can have one job per drive? 
>> 
>>If such restriction exists, we need to remove it anyways. 

Yes, I think we can have 1 job per drive at the same time. (as we can only do 1 block-job-complete). 

>>If such restriction exists, we need to remove it anyways. 
Out of my competence ;) But maybe you known how to do this in qemu code. 


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

De: "Dietmar Maurer" <dietmar at proxmox.com> 
À: "Alexandre DERUMIER" <aderumier at odiso.com> 
Cc: "Michael Rasmussen" <mir at datanom.net>, pve-devel at pve.proxmox.com 
Envoyé: Lundi 21 Janvier 2013 08:52:29 
Objet: RE: [pve-devel] vm migration + storage migration with ndb workflow notes 

> >>Would it be possible to do all work using 'qmp' commands. We can simply 
> start the VM 
> >>if the VM is offline (paused state). I currently do that with the new backup 
> code. 
> >> 
> >>I guess that would make the whole thing much simpler? 
> Not sure about that, 
> About my templates patches, I can see one better thing with qemu-img, we 
> can copy multiple time at once. 
> with drive-mirror, we can only do 1 drive-mirror job at once. 

Are you sure? Usually you can have one job per drive? 

If such restriction exists, we need to remove it anyways. 
_______________________________________________ 
pve-devel mailing list 
pve-devel at pve.proxmox.com 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel