[pve-devel] ceph: clone vm

[Alexandre DERUMIER]

>>it adds snapshot support but not linked clones? or do i miss something. 

this add support to clone from a snapshot from a vm. (at not the default __base__ snapshot, from base-xxx templates ).

I don't have tested it, but this is almost all we need. (maybe some checks to add in qemuserver.pm, but nothing too big)
----- Mail original ----- 

De: "Stefan Priebe" <s.priebe at profihost.ag> 
À: "Alexandre DERUMIER" <aderumier at odiso.com>, "Michael Rasmussen" <mir at datanom.net> 
Cc: pve-devel at pve.proxmox.com 
Envoyé: Mercredi 5 Février 2014 20:50:34 
Objet: Re: [pve-devel] ceph: clone vm 

it adds snapshot support but not linked clones? or do i miss something. 

Stefan 

Am 05.02.2014 16:06, schrieb Alexandre DERUMIER: 
>>> Cloning from snapshot is also an opportunity with ZFS. 
> 
> yes, sure, for any storage which support snasphot. 
> 
> It's not a big work, as we have already almost all subroutines in storage plugins 
> 
> 
> 
> something like this should work (don't have tested it ;) 
> 
> 
> qemuserver.pm 
> -------------- 
>>> sub clone_disk { 
>>> my ($storecfg, $vmid, $running, $drivename, $drive, $snapname, 
>>> $newvmid, $storage, $format, $full, $newvollist) = @_; 
>>> 
>>> my $newvolid; 
>>> 
>>> if (!$full) { 
>>> print "create linked clone of drive $drivename ($drive->{file})\n"; 
>>> $newvolid = PVE::Storage::vdisk_clone($storecfg, $drive->{file}, $newvmid); 
> 
> sub clone_disk { 
> my ($storecfg, $vmid, $running, $drivename, $drive, $snapname, 
> $newvmid, $storage, $format, $full, $newvollist) = @_; 
> 
> my $newvolid; 
> 
> if (!$full) { 
> print "create linked clone of drive $drivename ($drive->{file})\n"; 
> $newvolid = PVE::Storage::vdisk_clone($storecfg, $drive->{file}, $newvmid, $snapname); 
> 
> 
> storage.pm 
> ---------- 
>>> sub vdisk_clone { 
>>> my ($cfg, $volid, $vmid) = @_; 
>>> 
>>> my ($storeid, $volname) = parse_volume_id($volid); 
>>> 
>>> my $scfg = storage_config($cfg, $storeid); 
>>> 
>>> my $plugin = PVE::Storage::Plugin->lookup($scfg->{type}); 
>>> 
>>> activate_storage($cfg, $storeid); 
>>> 
>>> # lock shared storage 
>>> return $plugin->cluster_lock_storage($storeid, $scfg->{shared}, undef, sub { 
>>> my $volname = $plugin->clone_image($scfg, $storeid, $volname, $vmid); 
>>> return "$storeid:$volname"; 
>>> }); 
>>> } 
> 
> 
> sub vdisk_clone { 
> my ($cfg, $volid, $vmid, $snap) = @_; 
> 
> my ($storeid, $volname) = parse_volume_id($volid); 
> 
> my $scfg = storage_config($cfg, $storeid); 
> 
> my $plugin = PVE::Storage::Plugin->lookup($scfg->{type}); 
> 
> activate_storage($cfg, $storeid); 
> 
> # lock shared storage 
> return $plugin->cluster_lock_storage($storeid, $scfg->{shared}, undef, sub { 
> my $volname = $plugin->clone_image($scfg, $storeid, $volname, $vmid, $snap); 
> return "$storeid:$volname"; 
> }); 
> } 
> 
> 
> storage plugin 
> -------------- 
>>> sub clone_image { 
>>> my ($class, $scfg, $storeid, $volname, $vmid) = @_; 
>>> 
>>> my $snap = '__base__'; 
>>> 
>>> my ($vtype, $basename, $basevmid, undef, undef, $isBase) = 
>>> $class->parse_volname($volname); 
>>> 
>>> die "clone_image onyl works on base images\n" if !$isBase; 
>>> 
>>> my $name = &$find_free_diskname($storeid, $scfg, $vmid); 
>>> 
>>> warn "clone $volname: $basename to $name\n"; 
>>> 
>>> my $newvol = "$basename/$name"; 
>>> 
>>> my $cmd = &$rbd_cmd($scfg, $storeid, 'clone', &$add_pool_to_disk($scfg, $basename), '--snap', $snap, &$add_pool_to_disk($scfg, $name)); 
>>> run_command($cmd, errmsg => "rbd clone $basename' error"); 
>>> 
>>> return $newvol; 
>>> } 
> 
> sub clone_image { 
> my ($class, $scfg, $storeid, $volname, $vmid, $snapshot) = @_; 
> 
> my $snap = '__base__'; 
> $snap = $snapshot if $snapshot; 
> 
> my ($vtype, $basename, $basevmid, undef, undef, $isBase) = 
> $class->parse_volname($volname); 
> 
> die "clone_image onyl works on base images\n" if !$isBase && !$snapshot; 
> 
> my $name = &$find_free_diskname($storeid, $scfg, $vmid); 
> 
> warn "clone $volname: $basename to $name\n"; 
> 
> my $newvol = "$basename/$name"; # Maybe find a way to reference snasphot like "$basename@$snasphot/$name" 
> 
> my $cmd = &$rbd_cmd($scfg, $storeid, 'clone', &$add_pool_to_disk($scfg, $basename), '--snap', $snap, &$add_pool_to_disk($scfg, $name)); 
> run_command($cmd, errmsg => "rbd clone $basename' error"); 
> 
> return $newvol; 
> } 
> 
> 
> 
> ----- Mail original ----- 
> 
> De: "Michael Rasmussen" <mir at datanom.net> 
> À: pve-devel at pve.proxmox.com 
> Envoyé: Mercredi 5 Février 2014 15:39:23 
> Objet: Re: [pve-devel] ceph: clone vm 
> 
> On 02-05-2014 14:50, Alexandre DERUMIER wrote: 
>> 
>> Maybe could we introduce this feature (for ceph,sheepdog,..), cloning 
>> from a snapshot. 
>> 
> Cloning from snapshot is also an opportunity with ZFS. 
>