Moving Performance VM vs LXC

[leex12]

I have a simple homelab four node cluster. Running ceph for VM and LXC storage. I have two seperate pools at the moment as I moving stuff around. One pool is HD based and the other SSD.

On the same node from HDD to SSD poolstakes massively different times. VM 100gb drive shifted over in 15mins LXC 100gb took 44mins.

Does that mirror other peoples experience?

PLEASE PLEASE .. can we have a ETA, progress bar or something anything for LXC moves

 

[datschlatscher]

While copying block devices for VMs is rather straightforward and quick because it is in a way just like copying one big, contiguous file, LXCs are cloned based on their whole file hierarchy. That is why it can take much longer to copy/move a container of similar size, because the disks have to read and write each file separately.

PLEASE PLEASE .. can we have a ETA, progress bar or something anything for LXC moves

The problem is, it is really hard to know for sure how many files and how much time is left when copying or moving large amounts of single files. One would have to come up with a good way to implement such a progress indicator to work seamlessly with our task viewer, or do it in some other way first. But as far as I know, nobody has taken up that task yet.