[ovirt-users] Clone of VM and copy of disks doubts

[Gianluca Cecchi]

Hello,
I have a VM VM1 that I would like to clone.
It has 3 disks: 2 of them on storage domain SDA, 1 of them (size 80Gb) on a
second storage domain SDB

I want the new VM to have all the 3 disks on SDA, because SDB has not
enough space to accomodate the cloned disk

I power off VM1, select it and then "Clone VM". It seems I have no options
at all and that I can only specify the name of the new VM, while oVirt
would copy the VM with exactlly the same disks and so I receive error

Error while executing action:

VM2:

   - Cannot add VM. Low disk space on Storage Domain SDB.

Is this expected? In case it seems not very flexible.

So I create a snapshot of VM1 and then "Clone". Inside the "Resource
Allocation" section, the drop-down for the 3 disks seems not enabled, in
the sense that for every disk it contains only the name of the origin SD,
so for 2 disks it contains SDA and no other SD to be chosen, for 1 disk it
contains SDB and no other SD to be chosen.
See here:
https://drive.google.com/file/d/0BwoPbcrMv8mvdFd0RFMzRXhmRzQ/view?usp=sharing
Can anyone replicate this in 4.1.5 to see if it is a problem of mine?

So as a last resource I decide to copy disk by disk.
Apparently my VM1 disk in SDB is preallocated: at least this is what I see
if I select VM1, Snapshots, select the "Current" line (it is the only oe as
therer are no snapshots), "DIsks" in right subpane
See here for screenshot:
https://drive.google.com/file/d/0BwoPbcrMv8mvclV5bDdsdE9Ubmc/view?usp=sharing

Any more comfortable and correct way from web admin gui to see disk type?

When I copy the disk it seems it tries to create a thin provisioned disk
and it takes about 4 hours to copy 80Gb of disks from SSD to SAS SD with
very few I/O activity on hypervisor in the mean time (I have only one node
in this test)
when I would expect around 40minutes with a throughput of at elast 40MB/s
with a sort of dd as the cemu-img command should be... instead the command
I see from command line is kind of
 /usr/bin/qemu-img convert -p -t none -T none -f qcow2
/rhev/data-center/59b7af54-0155-01c2-0248-000000000195/70411599-c19f-4a6a-80b5-abdeffc1616e/images/3d357b96-57e1-47bd-be65-558252a591d3/77ba5e2b-dfbe-4801-b5ec-dc9ffae2c481
-O qcow2 -o compat=1.1
/rhev/data-center/mnt/blockSD/fad05d79-254d-4f40-8201-360757128ede/images/73292653-f5d9-49ae-8132-f17d288bf822/deb4b841-95cd-4523-9523-f2ffdbb9b4a9


Also, if get this VM and create a snapshot of it I see this:
- starting point all the 3 disks seem preallocated (same method as above)
- after creating a snapshot, if I select the "Current" line the 3 disks are
all thin-provision, if I select the line of the taken snapshot instead the
disks are all preallocated.

Any comments or suggestions about the points above?

Thanks,
Gianluca
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20170925/1d476093/attachment.html>