Daniel P. Berrange wrote:
> On Tue, Aug 12, 2008 at 08:21:15PM -0400, Cole Robinson wrote:
>> The attached patch adds virtinst support for building
>> logical pools and volumes. Sort of. This will allow
>> creating new volumes, but not creating pools (aka
>> volume groups) from a specified set of physical disks.
>> I believe libvirt supports this, but I haven't tried it
>> yet.
>
> Not a huge issue - using a pre-existing volume group is by
> far the most important use case here.
>
>> However this does allow the user to point at an existing
>> lvm volume group and have it recognized as a storage pool
>> which will be by far the common case. The nice thing
>> about this is that it will all become immediately available
>> in the virt-manager UI (posted previously).
>
> Looks good.
>
>> One thing I've encountered with this though: does a volume's
>> target path have any significance upon creation? File volumes
>> just seem to create a file based on the passed 'name', not
>> target path. Disk volumes ignore both when creating. Logical
>> vols also only use the name, but try to validate the target
>> path after creation, which we shouldn't need passed (the result
>> should simply be vgpath + "/" + volname). So it seems like
>> target path has no real meaning when defining a volume. Am
>> I missing something?
>
> That is correct - the target path for a volume is filled in
> according to the policy of the storage pool, often based off
> the name of the storage volume + the target path of the
> storage pool
>
> Daniel
>

Okay, I've committed this.

http://hg.et.redhat.com/virt/applications/virtinst--devel?cs=95dfaa9a29a6

Thanks,
Cole

_______________________________________________
et-mgmt-tools mailing list
et-mgmt-tools@redhat.com
https://www.redhat.com/mailman/listinfo/et-mgmt-tools