[Users] Questions about oVirt and possible API bug

Andrey Falko afalko at salesforce.com
Thu Apr 5 17:58:41 UTC 2012


> -----Original Message-----
> From: Itamar Heim [mailto:iheim at redhat.com]
> Sent: Wednesday, April 04, 2012 11:13 PM
> To: Andrey Falko
> Cc: users at oVirt.org
> Subject: Re: [Users] Questions about oVirt and possible API bug
>
> On 04/05/2012 01:30 AM, Andrey Falko wrote:
> > Hi everyone,
> >
> > I have one more question:
> >
> > I want to use local storage, but I want to register the same template on each node. I am working on a mechanism right
> > now where I detach and remove the Export domain, add it for a target dc/cluster/host, import the template,
> > deactivate+detach+remove the export domain and repeat for the next host. Is this mechanism the only choice I have? Is
> > there a better way?
>
> you need the template on each DC (until you would be able to mix types of storage domains and use a shared NFS for the
> template for example).
> this feature should allow you to import it more than once, but not implemented yet:
> http://www.ovirt.org/wiki/Features/ImportMoreThanOnce
>
> that hints your current solution will not work, since you cannot import the same object today more than once (even to
> another DC).
>
> since this is local storage, you can hack it a bit and just create a template with the disk on all DCs, then dd/rsync
> the real one to override the created disk in all local storage.
>

Good to know the feature is in the pipeline :). I was able to work around it by writing a script that does the
following:

a) copy origin OVF with dir structure in /master in the export domain
b) copy dir structure for associated image in /images
c) hard link origin image to the new image uuid dir
d) replace uuid in .meta file and uuids in new OVF
e) additionally, replace instanceId for VM network as well as storagePoolId (not sure if storagePoolId replacement was
necessary

Thanks,
Andrey



More information about the Users mailing list