[pve-devel] [PATCH v2 storage 1/9] storage: add new find_free_volname

Aaron Lauterer a.lauterer at proxmox.com
Fri Aug 13 14:46:51 CEST 2021



On 8/12/21 2:50 PM, Fabian Ebner wrote:
> Am 06.08.21 um 15:46 schrieb Aaron Lauterer:
>> This new method exposes the functionality to request a new, not yet
>> used, volname for a storage.
>>
>> The default implementation will return the result from
>> 'find_free_diskname' prefixed with "<VMID>/" if $scfg->{path} exists.
>> Otherwise it will only return the result from 'find_free_diskname'.
>>
>> If the format suffix is added also depends on the existence of
>> $scfg->{path}.
>>
>> $scfg->{path} will be present for directory based storage types.
>>
>> Should a storage need to return a different volname, it needs to override
>> the 'find_free_volname' method.
>>
>> Signed-off-by: Aaron Lauterer <a.lauterer at proxmox.com>
>> ---
>>
>> v1 -> v2:
>> * drop exposed 'find_free_diskname' in favor of 'find_free_volname'
>> * dropped 'wants_fmt_suffix' as 'find_free_volname' now decides that itself
>>
>> rfc -> v1:
>> dropped $add_fmt_suffix parameter and added the "wants_fmt_suffix"
>> helper method in each plugin.
>>
>>   PVE/Storage.pm        |  9 +++++++++
>>   PVE/Storage/Plugin.pm | 11 +++++++++++
>>   2 files changed, 20 insertions(+)
>>
>> diff --git a/PVE/Storage.pm b/PVE/Storage.pm
>> index c04b5a2..c38fe7b 100755
>> --- a/PVE/Storage.pm
>> +++ b/PVE/Storage.pm
>> @@ -203,6 +203,15 @@ sub storage_can_replicate {
>>       return $plugin->storage_can_replicate($scfg, $storeid, $format);
>>   }
>> +sub find_free_volname {
>> +    my ($cfg, $storeid, $vmid, $fmt) = @_;
>> +
>> +    my $scfg = storage_config($cfg, $storeid);
>> +    my $plugin = PVE::Storage::Plugin->lookup($scfg->{type});
>> +
>> +    return $plugin->find_free_volname($storeid, $scfg, $vmid, $fmt);
>> +}
>> +
>>   sub storage_ids {
>>       my ($cfg) = @_;
>> giff --git a/PVE/Storage/Plugin.pm b/PVE/Storage/Plugin.pm
>> index b1865cb..e043329 100644
>> --- a/PVE/Storage/Plugin.pm
>> +++ b/PVE/Storage/Plugin.pm
>> @@ -685,6 +685,17 @@ sub find_free_diskname {
>>       return get_next_vm_diskname($disk_list, $storeid, $vmid, $fmt, $scfg, $add_fmt_suffix);
>>   }
>> +sub find_free_volname {
>> +    my ($class, $storeid, $scfg, $vmid, $fmt) = @_;
>> +
>> +    my $diskname = $class->find_free_diskname($storeid, $scfg, $vmid, $fmt, exists($scfg->{path}));
>> +
>> +    if (exists($scfg->{path})) {
> 
> More of a nit, but I don't like the usage of exists here, for two reasons:
> 
> 1. The main reason: it's inconsistent with all the other checks for $scfg->{path} that are used throughout the file.
> 
> 2. It's more likely to break. Consider some code I've made up, but similar code might exist somewhere in our code base:
>      ...
>      my $path = get_path_from_config($cfg, $storeid);
>      my $scfg = {
>      path => $path,
>      content => $content,
>      ...
>      };
> or
>      my $start = $scfg->{path}[0];
>      # auto-vivication means the 'path' key now always exists (with undef value if it was undef before)
>      if (defined($start) && $start eq '/') {
>      ...
>      }
> 
> Using exists here makes it necessary to check that such code is not used for $scfg, currently and in the future.

Okay I see your point. So following the regular approach throughout the pve-storage repo should be okay. That is, evaluating $scfg->{path} directly.

> 
>> +    return "${vmid}/$diskname";
>> +    }
>> +    return $diskname;
>> +}
>> +
>>   sub clone_image {
>>       my ($class, $scfg, $storeid, $volname, $vmid, $snap) = @_;
>>





More information about the pve-devel mailing list