Large-scale migration from VMWare to Proxmox (Ceph)

[breakaway9000]

Hi,

I'm working on a project right now where I've got a number of vmware VMs (approx 50) taking up approx 5-8 TB of space. They're running on a VMWare cluster.

I've also got a 3-node proxmox cluster setup with Ceph up and running - ready to migrate off VMware to Proxmox/Ceph.

I've done this migration many times before but never to Ceph storage (Always NFS or just plain old SSD storage).

What's the best way to accomplish this? My research shows a few people have migrated the .vmdk across to proxmox storage first (e.g. a SSD or HDD installed in the host, partitioned & formatted to ext3/4/zfs), then moved it across into ceph by using the proxmox interface.

Is there a way to go direct from VMWare to ceph? It'd save a lot of time.

Thanks.

 

[aderumier]

Hi,

I'm working on a project right now where I've got a number of vmware VMs (approx 50) taking up approx 5-8 TB of space. They're running on a VMWare cluster.

I've also got a 3-node proxmox cluster setup with Ceph up and running - ready to migrate off VMware to Proxmox/Ceph.

I've done this migration many times before but never to Ceph storage (Always NFS or just plain old SSD storage).

What's the best way to accomplish this? My research shows a few people have migrated the .vmdk across to proxmox storage first (e.g. a SSD or HDD installed in the host, partitioned & formatted to ext3/4/zfs), then moved it across into ceph by using the proxmox interface.

Is there a way to go direct from VMWare to ceph? It'd save a lot of time.

Thanks.

proxmox 5.0 have a new command "qm importdisk" to import vmdk file directly to ceph (or any other storage).
proxmox 5.1 should have qm importovf, to import ovf without vm config + disk. (code has been commited some days ago)

 

[breakaway9000]

Unfortunately due to no stable release of ceph being available on 5.0 or 5.1, I can't quite upgrade just yet. Is there a solution that doesn't involve using those commands? I don't have them on 4.4.

 

[kobuki]

The way I usually do it is creating the target VM with appropriate config and disks, then use qemu-nbd to present the vmdk as a block device and just dd it over to the new VM's disk. It can be easily automated in scripting. I haven't used ceph yet but I suppose it provides a logical block device for this kind of procedure.

 

[Daniel Keller]

if your vmdk files on an nfs share, you can use them in proxmox directly for your VMs and then move them to ceph storage with move disk.

or you use dd like this

dd if=/dev/somedvice | pv | ssh root@remote-server 'rbd --image-format 2 import - pool/webserver'

 

[kobuki]

dd if=/dev/somedvice | pv | ssh root@remote-server 'rbd --image-format 2 import - pool/webserver'

Wouldn't it rather be:

dd if=/path/to/device | pv -s <size of device> | ssh root@remote-server 'rbd --image-format 2 import - pool/webserver'

Device is supposedly the product of using qemu-rbd, right?

EDIT: fixed my suggestion

 

[Daniel Keller]

no somedevice woud be a mounted vmdk or the with qemu-nbd presented vmdk

pv shows only the progress of the operation

the same like here only different input http://wiki.skytech.dk/index.php/Ceph_-_howto,_rbd,_lvm,_cluster#Import_LVM_image_into_rbd

 

[kobuki]

no somedevice woud be a mounted vmdk or the with qemu-nbd presented vmdk

pv shows only the progress of the operation

Yes, I fixed my reply in the meantime realising that. But pv needs the size when fed through stdin.