[pve-devel] [PATCH 2/7] use qemu_drive_mirror_monitor in live full clone

[Alexandre DERUMIER]

>>I wonder if we should give the clone_disk() - which is called above in a
>>loop - a parameter telling it to do this. Not only would save us a
>>query-block-job call in the cases where we qemu_drive_mirror() isn't
>>called, but since this is the $running case, I'm worried about too many
>>parallel mirror jobs hitting a networking bottle-neck and failing to
>>complete where a series of single-disk mirrors would succeed.

So, begin to mirror first disk, wait until it reach 100% but don't block-job-complete,
then mirror second disk, ... ?

I don't have checked network bandwith when multiple job are running in parallel.
Not sure if we should add an option to define number of parallel job ?




----- Mail original -----
De: "Wolfgang Bumiller" <w.bumiller at proxmox.com>
À: "aderumier" <aderumier at odiso.com>
Cc: "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Jeudi 20 Octobre 2016 09:53:14
Objet: Re: [pve-devel] [PATCH 2/7] use qemu_drive_mirror_monitor in live full clone

On Thu, Oct 20, 2016 at 02:35:11AM +0200, Alexandre Derumier wrote: 
> Signed-off-by: Alexandre Derumier <aderumier at odiso.com> 
> --- 
> PVE/API2/Qemu.pm | 1 + 
> 1 file changed, 1 insertion(+) 
> 
> diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm 
> index ad7a0c0..98865f6 100644 
> --- a/PVE/API2/Qemu.pm 
> +++ b/PVE/API2/Qemu.pm 
> @@ -2415,6 +2415,7 @@ __PACKAGE__->register_method({ 
> PVE::QemuConfig->write_config($newid, $newconf); 
> } 
> 
> + PVE::QemuServer::qemu_drive_mirror_monitor($vmid,$newid) if $running && $param->{full} && !$snapname; 

I wonder if we should give the clone_disk() - which is called above in a 
loop - a parameter telling it to do this. Not only would save us a 
query-block-job call in the cases where we qemu_drive_mirror() isn't 
called, but since this is the $running case, I'm worried about too many 
parallel mirror jobs hitting a networking bottle-neck and failing to 
complete where a series of single-disk mirrors would succeed. 

> delete $newconf->{lock}; 
> PVE::QemuConfig->write_config($newid, $newconf); 
> 
> -- 
> 2.1.4 
> 
> _______________________________________________ 
> pve-devel mailing list 
> pve-devel at pve.proxmox.com 
> http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel