using lvconvert for mirroring

I have been playing with lvconvert but I can't quite figure out what it is doing. If I start with this:

Debian# lvs -o +devices
LV VG Attr LSize Origin Snap% Move Log Copy% Devices
lvol0 vg1 -wima- 7.81G /dev/hde2(0)

Then run:

Debian# lvconvert -m1 vg1/lvol0

I get this:

Debian# lvs -o +devices
LV VG Attr LSize Origin Snap% Move Log Copy% Devices
lvol0 vg1 mwima- 7.81G lvol0_mlog 2.75 lvol0_mimage_0
(0),lvol0_mimage_1(0)

So my first question is how do I find out where the mirror really is? I have also tried specifying the physical volume at the end of the command to put the mirror on a different pv but then I get this error message:

Debian# lvconvert -m1 vg1/lvol0 /dev/hdg1
Not enough PVs with free space available for parallel allocation. Consider --alloc anywhere if desperate.

Thank you for any insight.

Try using the additional '-a' argument (e.g. 'lvs -a -o +devices). That should tell you more.

The reason you are getting errors with the last convert mentioned is because you didn't specify enough devices. Mirroring (with disk logging) requires at least 3 devices - two for the actual mirror images and one for the log.

There are some pending enhancements to the allocation code that allow the log device to be placed on the same physical disk as a mirror device - thus only requiring two devices. I'm not sure when the allocation changes will go in.

brassow