Hello,
is it already possible to move vm's from one datastore to another without downtime (storage vmotion
) with vCloud Director. If not yet - is it planned to implement this ? This is not a feature for the Org-Admin, but for the vCloud Admin, especially when storage-systems are going to replaced. And don't tell me to use some storage-virtualisation product. Even those have to be replaced someday ![]()
This would of cource mean, that the Org-Admin would see multiple datastores while migrating the vapps. One thing to consider is of course how to handle fast-provioned vm's. Consolidating is no real option, because the vm has to be powered off (at least from gui). And of course it is a one-way solution...
Thanks
You should be able to use Storage vMotion directly *BUT* there are some *gotchas* you need to be aware of :
1) You *need* to storage vMotion the VM to a datastore present into your PvDC
2) If you are using Fast Provisionning (Linked Clones) the Storage vMotion, will not keep it as the source format, and *WILL* consolidate the VM.
If you want to keep the disk format, use the vCloud API and the *relocate* api call (it will take care of everything needed to keep it as a linked clone).
Note : I'm talking about 1.5.x, as you have SDRS & Storage profiles in 5.1, you can manage differently with SDRS Maintenance mode>
Hope this helps,
You should be able to use Storage vMotion directly *BUT* there are some *gotchas* you need to be aware of :
1) You *need* to storage vMotion the VM to a datastore present into your PvDC
2) If you are using Fast Provisionning (Linked Clones) the Storage vMotion, will not keep it as the source format, and *WILL* consolidate the VM.
If you want to keep the disk format, use the vCloud API and the *relocate* api call (it will take care of everything needed to keep it as a linked clone).
Note : I'm talking about 1.5.x, as you have SDRS & Storage profiles in 5.1, you can manage differently with SDRS Maintenance mode>
Hope this helps,
Hi,
that means if we are using 5.1,we will be able to put datastores in maintenance mode, and sdrs will do all the rest ? Storage vMotion VMs away from affected datastore, take care of linked clones, no Downtime of vms, etc ?
I haven't had the time until now to go into details with SDRS, but it looks like i should do so. Will it also be possible to remove datastore completely when it is in maintenance mode, and add new datastore before/after that. That would be perfect to relocate vm's if storage systems are going to be retired...
Thank you
From beta testing and what I remember I don't think SDRS is Link-clone aware in 5.1. I know it's not in 5.0 and have done a lot of testing around this as we are migrating from Lab Manager to vCD.
I just upgraded this week to VCD/vCenter/ESXi 5.1 and haven't had a chance to verify this yet though. I'm hoping to test this next week.
I did a quick test with the REST API. That did work perfect. The machine was tranferred while running and it was still a linked clone after relocating. The parent image was located in the same directory as the relocated vm. I have to check what happens when relocating vm's that use the same parent image to the same "new" datastore. Will another parent be created in second the vm's directory, or will the parent of the first vm be used ?
We were also able to relocate back to the originating datastore. In that case the orignal parent was referenced again.
Anyhow, that did what we want to achieve. Next step would be to do it with vCenter Orchestrator. Since we are not replacing storage systems very often, that will be a solution we can live with very good.
Thank you guys...
