Hello Sam,
Welcome to vSAN.
"I wanted to know if i would create instead 2 Disk groups on each host would improve performance."
Yes, likely significantly as you will have double the cache to capacity ratio you have now.
Whether this will be significantly noticeable depends on whether you are currently pushing the current configuration to near the limits of its capabilities.
"if its possible to do this with now downtime."
Yes, it is possible without downtime and quite straight-forward:
1. Ensure you have enough space to comfortably evacuate one Capacity-tier disk - as it is a 3-node cluster, assuming everything is stored as FTT=1,FTM=RAID1 (the default) you would need to have enough available space on a single node (as that is the only valid location for placement of components from the disk you are evacuating because it can't move them to the other nodes as that would violate the Storage Policy). If you don't have enough space (on each node) to remove a disk with 'Full Data Evacuation' option then you can use 'Ensure Accessibility' option though note that some data will be FTT=0 until it is resynced back to FTT=1 so ensure you have good backups.
2. Remove 1 Capacity-tier disk from a Disk-Group with whichever option is viable.
3. Create second Disk-Group using your unused Cache-tier SSD and the newly blanked Capacity-tier disk.
4. Proceed to the same for the next 2 disks on this node, adding them to the new Disk-Group - if you didn't have enough space for it with the 1st disk, you should actually be able to remove each one after the first with 'Full Data Evacuation' option as the data will be able to be moved to the new Disk-Group.
5. Proceed to do the same for the other nodes.
"Also would having 2 disk groups and same amount of capacity disk would that leave me with same datastore size?"
Yes. The only time this may not hold true would be if deduplication was enabled (as by the law of odds, you would get a lower dedupe ratio with less data on each Disk-Group).
Bob
