Have SATA and SAS SSD's as capacity in VSAN servers.

[jpab-01]

Greetings gurus,

I have a situation where we have 2 servers with SATA as capacity drives and I would like to add additional storage using SAS in the 2 servers. Is this possible and supported?

Thank you in advance for your assistance.

[TheBobkin]

@jpab-01, mixing different models of SSDs in a Disk-Group/node is supported as long as the new devices are on the vSAN HCL and rated for the intended purpose (e.g. All-Flash Capacity-tier).

That being said, you should ideally be ensuring devices you add don't vary drastically in performance spec from the rest of the hardware (or at the least are better not worse performing).

[jpab-01]

Hi The BobKin,

thank you in getting back to me regarding this question. We have created a new disk group and added the SAS disks to it. The VSAN automatically assigned the new disk group to the total storage of the VSAN. 

Is this acceptable?

Regards

[TheBobkin · Accepted solution]

Hi jpab-01,

So, it being able to create Disk-Groups and write vSAN partitions to the devices doesn't really test much other than the devices are detected by ESXi, didn't have partitions and can be written to.

I would advise as per my last comment to validate and compare the disk models in use of the previous and new disks to ensure they are all on the vSAN HCL, and have been certified for the use you are using them for (e.g. All-Flash Cache-tier, All-Flash Capacity-tier) and that the new devices are same or better performance than the previous disks (e.g. so that they are not going to decrease the performance of the cluster).

It would also be advisable to compare the cluster performance before and after adding these but only after some data has been moved to these disks - this doesn't happen automatically unless you have Capacity-tier disks >80% used or have disks with >30% difference between lowest and highest used disk AND have automatic rebalance configured - once a reasonable amount of data has been moved to the new disks and resyncs are not ongoing, you should look at a typical days performance at the cluster level before and after the hardware change to observe the latency and throughput to see is there a marked difference (Cluster > Monitor > vSAN > Performance > VM Frontend and Backend being the most useful).