[Users] theoretical scenario

Itamar Heim iheim at redhat.com
Wed Dec 12 23:47:30 UTC 2012


On 12/11/2012 09:47 PM, Jonathan Horne wrote:
> ok, so the process seems pretty obtuse, so let me run this by the list
> here and maybe I'm just not doing it the best way.
>
> 1) i connected an export domain called EXPORTNFS.  the nfs share is
> physically on my management server (as is iso)
> 2) exported 1 VM.
> 3) on the mgmt server, i run "tar zxcf /root/test.tgz
> /opt/export/[UUID]/images /opt/export/[UUID]/master"
> 4) for good measure, i use ovirt web gui to delete the exported VM from
> the export
> 5) run the command 'engine-image-uploader –u admin at internal upload
> test.tgz –v –e EXPORTNFS
> 6) after a while, it completes and i see the VM back in the export
> storage tab under VM Import.
>
> i didn't run the vm yet, but I'm assuming thats a successful export and
> import.
>
> so, it seems if i wanted to export each and every VM from one separate
> ovirt to another, i would export each VM, tar-gzip it, clear it from the
> webgui, and start the next one?   IE.. i mean it looks like
> engine-image-uploader would be pretty unhappy with finding more than one
> set of VM UUIDs inside?

apart of Vered's simpler solution, a small note that:
1. engine-image-uploader can handle uploading the same image more than 
once and generating uuid's
2. however, i think this capability is deprecated, since engine has 
recently added support to import a VM which UUID is already imported, 
allowing to create a new VM from the imported one.



More information about the Users mailing list