When i say forever, it means many many hours, nothing happens.
I inadvertently created a couple of VMs using my NAS network drive instead of local-lvm that I would like to move these back to local-LVM
No go through the GUI (hangs forever) and no go through command line either using:
qm move_disk VMID source_disk target_storage
which results in command alot like the one that is launched by the GUI:
'/usr/bin/qemu-img convert -p -n -f qcow2 -O raw /mnt/pve/ProxMoxShare/images/110/vm-110-disk-0.qcow2 zeroinit:/dev/pve/vm-110-disk-0'
I observed that while this command tries to run (and do nothing, just sits there forever) there is no activity on this network leg at all on my managed switch and also by doing a pvesm status, my Network Share becomes inactive. Come back active when i stop the command from the gui.
Network share is on a Qnap that has a capacity of 800-900MB/s transfer speed so this is not the bottleneck.
I tried to fool the system (during one full day) trying to bring my image to "the box"
1. adding a USB key and format it with proxmox filing system
2. upload a qcow2 image (by going around the GUI that won't accept anything else than iso file so temporarily renaming it to iso)
3. Added a new lvm disk on the VM
4. back to cmd line, rename iso to qcow2 and initiated the above command (convert) over that disk -> It worked but the result Won't boot
5. It tried to go back to my initial network share, no boot anymore. I would guess that by mounting it again, it was too stoopid and created an empty share instead of reusing the one already there
6. Fortunately i had backed up the original qcow2 image, so i copy it over the network share, this won't boot again
7. I still have that iso renamed to qcow2 into /mnt/pve/USB/template/iso
but regardless what i try, it doesn't seem to go. I don't understand why it works for other ppl.
I read a thread saying that the default option for network shares was put to async_io uring and that could cause the hang-up. But i also tried to change that successively to 'native' and 'threads' - the result is the same. GUI will hand forever.
Remember when GUI hangs, pvesm status becomes inactive and as soon as i stop the GUI it comes back.
Also it seems that if i try to access this network share from any other node (i have 4 total) when it is in this state, it is "kaput" ! untill i stop the initiated move in the origin GUI.
I will go out on a limb here and say "maybe" this function need refinement. It is bizarre that it simply hangs up forever without giving any error.
But i saw some ppl on youtube having it work. Me it's a no go. I think i will re-create all the VMs from scratch but it is very limiting to be 'stucked forever' where you initially put the disk, becaus 1 error and you are doomed.
I inadvertently created a couple of VMs using my NAS network drive instead of local-lvm that I would like to move these back to local-LVM
No go through the GUI (hangs forever) and no go through command line either using:
qm move_disk VMID source_disk target_storage
which results in command alot like the one that is launched by the GUI:
'/usr/bin/qemu-img convert -p -n -f qcow2 -O raw /mnt/pve/ProxMoxShare/images/110/vm-110-disk-0.qcow2 zeroinit:/dev/pve/vm-110-disk-0'
I observed that while this command tries to run (and do nothing, just sits there forever) there is no activity on this network leg at all on my managed switch and also by doing a pvesm status, my Network Share becomes inactive. Come back active when i stop the command from the gui.
Network share is on a Qnap that has a capacity of 800-900MB/s transfer speed so this is not the bottleneck.
I tried to fool the system (during one full day) trying to bring my image to "the box"
1. adding a USB key and format it with proxmox filing system
2. upload a qcow2 image (by going around the GUI that won't accept anything else than iso file so temporarily renaming it to iso)
3. Added a new lvm disk on the VM
4. back to cmd line, rename iso to qcow2 and initiated the above command (convert) over that disk -> It worked but the result Won't boot
5. It tried to go back to my initial network share, no boot anymore. I would guess that by mounting it again, it was too stoopid and created an empty share instead of reusing the one already there
6. Fortunately i had backed up the original qcow2 image, so i copy it over the network share, this won't boot again
7. I still have that iso renamed to qcow2 into /mnt/pve/USB/template/iso
but regardless what i try, it doesn't seem to go. I don't understand why it works for other ppl.
I read a thread saying that the default option for network shares was put to async_io uring and that could cause the hang-up. But i also tried to change that successively to 'native' and 'threads' - the result is the same. GUI will hand forever.
Remember when GUI hangs, pvesm status becomes inactive and as soon as i stop the GUI it comes back.
Also it seems that if i try to access this network share from any other node (i have 4 total) when it is in this state, it is "kaput" ! untill i stop the initiated move in the origin GUI.
I will go out on a limb here and say "maybe" this function need refinement. It is bizarre that it simply hangs up forever without giving any error.
But i saw some ppl on youtube having it work. Me it's a no go. I think i will re-create all the VMs from scratch but it is very limiting to be 'stucked forever' where you initially put the disk, becaus 1 error and you are doomed.