online disk resize

[dimaandg]

No need to restart VM for disk resize, all OSs can understand that disk was resized. Just need to use qemu-monitor-command to inform OS about it.

example virsh qemu-monitor-command vm-2463 --hmp "block_resize drive-virtio-disk1 "

GOGS Ticket: https://docs.greenitglobe.com/gig/proj_russia_digital_energy/issues/348

[grimpy]

root@ovs1:~# virsh qemu-monitor-command vm-13 --hmp "block_resize drive-virtio-disk0 16106127360"
Could not resize: Operation not permitted

The storage system we use undernead does not seem to support this. Upstream ticket with FR was created. http://support.openvstorage.com/servicedesk/customer/portal/1/OVSSUP-324

[dmitrydelyukin]

I used this process for online disk resize and it works. I can send you step-by-step procedure if needed.

[grimpy]

@dmitrydelyukin this was with an OpenvStorage backend? If so please add step-by-step procedure so we can see if this helps.

[dmitrydelyukin]

# logging to storage node
# stor-0X
truncate -s <newsize> /mnt/<storagepoolname>/myvolume.raw
# need to find what cpu node is serving VM
# logging to cpu node
# cpu-XX
root@cpu-11:~# virsh list
 Id    Name                           State
—------------------------------------------------—
 110   vm-2463                        running
 root@cpu-11:~# virsh domblklist vm-2463 —details
Type       Device     Target     Source
—----------------------------------------------
network    cdrom      hdc        vm-2463/cloud-init-vm-2463
network    disk       vda        vm-2463/bootdisk-vm-2463
network    disk       vdb        volumes/volume_8198

root@cpu-11:~# virsh pool-list
 Name                 State      Autostart
—---------------------------------------—
root@cpu-11:~# virsh qemu-monitor-command vm-2463 —hmp "info block"
drive-ide0-1-0 (#block154): json:{"driver": "raw", "file": {"port": 26203, "host": "10.200.3.61", "driver": "openvstorage", "enable-ha": true, "volume": "vm-2463/cloud-init-vm-2463", "transport": "tcp"}} (raw, read-only)
    Removable device: locked, tray closed
    Cache mode:       writeback, direct

drive-virtio-disk0 (#block323): json:{"driver": "raw", "file": {"port": 26203, "host": "10.200.3.65", "driver": "openvstorage", "enable-ha": true, "volume": "vm-2463/bootdisk-vm-2463", "transport": "tcp"}} (raw)
    Cache mode:       writeback, direct

drive-virtio-disk1 (#block555): json:{"driver": "raw", "file": {"port": 26209, "host": "10.200.3.66", "driver": "openvstorage", "enable-ha": true, "volume": "volumes/volume_8198", "transport": "tcp"}} (raw)
    Cache mode:       writeback, direct

root@cpu-11:~# virsh qemu-monitor-command vm-2463 —hmp "block_resize drive-virtio-disk1 <newsize>"

[grimpy]

From within our python code

In [8]: dom.blockResize('vdb', 26777216, libvirt.VIR_DOMAIN_BLOCK_RESIZE_BYTES)

Result should be 0 if everything was successfull, I failure we return a 204 instead of a 200, this is an indication the caller of the API should restart his vm nevertheless. Possible reasons for live changing might be cause guest OS does not support it.

[john-kheir]

I changed the disk "test" to 21G but it is still 10G on the vm itself (vdb)