storage containers and images

Hi, Apart from doing it from prism element I don't know. I'm also interested to know if anyone has the answer/tips.

you can’t from PE either, if you select clone from image service there is no option to choose where the disk gets created

Same concern. It seems no way to have a simple solution. What we could do is either using CLI command or using PC to update disk container after VM created.

the option to do it in PC for a VM deployed via image is grayed out, so you have to use the acli 

That is how it works. Images uploaded via PC will come in the selfservicecontainer for all clusters that the image is uploaded to. 

Images uploaded via prism element you can choice a storagecontainer (during upload), not afterwards. 

All machines created from the image are placed in the storage container where the image resides. 

Now the fun part, try to switch over to use storage policies. So all vm's can stay in the selfservicestorage container. With storage policies you can change the compression/rf/encryption/etc when needed for that vm. So you never have to "storagemigrate” the vdisks again. (which needs to be done via cli and is not possible via PE/PC)

so you can change the compression\dedupe per VM rather than at the storagecontainer level?

​@Khoskin correct. But not dedupe ;) I've updated my post ;) 

ok we don’t use depupe or erasure coding anyway.  Are there any docs for best practice for what workloads benefit from offline\inline compression, dedupe and erasure coding?  We have a mixed workload of general purpose VM’s but a few fileservers and SQL boxes.

Hereby: https://portal.nutanix.com/page/documents/solutions/details?targetId=TN-2032-Data-Efficiency:TN-2032-Data-Efficiency#:~:text=Deduplication%20and%20compression%20generate%20dramatic,and%20workloads%20without%20affecting%20performance.

We recommend compression for almost all use cases except the following workloads:

• Encrypted datasets

• Compressed datasets (images, audio, and video)

We recommend using EC-X for write once, read many (WORM) workloads and other workloads with limited overwrites:

• Nutanix Objects

• Backups

• Archives

• File servers

• Log servers

• Email (depending on usage)

Don't use EC-X with the following workloads:

• Anything write or overwrite intensive

• VDI

We recommend deduplication for the following workloads:

• Full clones in VDI

• Persistent desktops

• P2V and V2V when using Hyper-V because Windows Offloaded Data Transfer (ODX) uses a full data copy