08.02.2016 20:45, Maxim Nestratov пишет:
08.02.2016 16:04, Olga Krishtal пишет:
> In-Reply-To:
>
> v2:
> - fixed memory leak
> - chenged the return value of all helper functions to 0/-1.
> Now check for success is smth like that: vir****Ploop() < 0
> - fixed some identation issues.
>
Overall looks good, except minor issues in some patches. Also, I
wonder if we should add some check to ensure ploop binary existence?
This would make user experience much better. For instance, instead of
unconditionally returning virStorageBackendCreatePloop in case of
ploop volumes in virStorageBackendGetBuildVolFromFunction we could
check for ploop external tool and report an error message that it
wasn't detected and it is required to be installed for PLOOP storage
pools.
Maxim Nestratov
--
libvir-list mailing list
libvir-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
Hmm. I thought over the idea of introducing a new storage pool type
PLOOP and I'm not sure that it is good. Actually it makes existing
storage pools like Directory unusable for ploop format without any
strong reason. Thus, I would tend to ask you to change the approach of
introducing a new pool storage type to using existing ones.