V2V import, qm importovf and ova archive

[Klug]

Hello.

I tried a V2V import based on the wiki steps.
https://pve.proxmox.com/wiki/Migration_of_servers_to_Proxmox_VE#CLI

I installed vmware's ovftool on a node and downloaded a VM from a vSphere Infrastructure (it works, you can connect to the vCenter and download through it, no need to connect to the ESX host).
I ended up with an OVA file (tar containing .ovh and .mf and .vmdk).

Then, based on the wiki steps, I tried to import the ova file.
This gives an error:

# qm importovf 191 ./downloaded_vm.ovf SSD
./downloaded_vm.ovf:1: parser error : Start tag expected, '<' not found
downloaded_vm.ovf
^

I found very little errors like this one on the net, about bad download (checksum error) or bad content of the ova.
But it seemed to me that "qm importovf" was not able to deal with the ova file.

So I untarred the ova (no error doing this) then relaunch the import, with the ovf file this time.
And it seems to work:

# qm importovf 191 ./downloaded_vm.ovf SSD
transferred: 0 bytes remaining: 161061273600 bytes total: 161061273600 bytes progression: 0.00 %
transferred: 1610612736 bytes remaining: 159450660864 bytes total: 161061273600 bytes progression: 1.00 %
transferred: 3221225472 bytes remaining: 157840048128 bytes total: 161061273600 bytes progression: 2.00 %
transferred: 4831838208 bytes remaining: 156229435392 bytes total: 161061273600 bytes progression: 3.00 %
transferred: 6442450944 bytes remaining: 154618822656 bytes total: 161061273600 bytes progression: 4.00 %
transferred: 8053063680 bytes remaining: 153008209920 bytes total: 161061273600 bytes progression: 5.00 %
transferred: 9663676416 bytes remaining: 151397597184 bytes total: 161061273600 bytes progression: 6.00 %
transferred: 11274289152 bytes remaining: 149786984448 bytes total: 161061273600 bytes progression: 7.00 %

So...

Is this an error in the wiki (you have to untar the file first)?
Or a disk space issue (there's not enough room for the untarred archive on the disk the ova is, I had to untar it on another disk)?
Or something else?

 

[Dominic]

Yes, the .ova has to be unpacked first. importovf basically reads the contents of the .ovf (the small description file) and then converts all the single .vmdk images that are placed in the folder next to the .ovf. It is interesting that you got an .ova file from the ovftool. I'll try to check this soon. Maaaybe this is because I used only one ESXi host in contrast to

downloaded a VM from a vSphere Infrastructure (it works, you can connect to the vCenter and download through it, no need to connect to the ESX host).

?

 

[Klug]

I actually name the file .ova on download but I could have used any other extension

But any name/extension I decide to use, I'll end up with a tar archive.

Shouldn't the wiki be upddated to reflect that (and that you have to untar it before use)?

 

[Dominic]

Shouldn't the wiki be upddated to reflect that (and that you have to untar it before use)?

Done

 

[Klug]

Thank you!

 

[Dominic]

Thank you for the hint!