Welcome to vSAN Communities!
"Some of the hosts have 2 disk groups and the remainder are only capable of a single disk group"
Are these out of available slots or some other reason?
How many nodes in the cluster and how many of each type?
While it may be technically supported to have non-homogeneous configurations, these are strongly advised against.
The possible outcome of having one node/Fault Domain with significantly varying amount of storage capacity available is that if you are using any SP (Storage Policy) with FTT=1, RAID1 (e.g. Default vSAN SP) then eventually it could end up significantly imbalanced e.g. with the lower-capacity nodes at 80-90% utilization while the higher-capacity node will only be at ~40-50% utilization. This is due to the fact that an FTT=1 RAID1 vSAN Object requires 2 nodes/fault-domains for the data components (+1 for tiny witness component), if component placement went ideally then potentially this *could work* e.g. always placing data-components on the higher-capacity node and alternating witness components on the other smaller-capacity nodes but there is no way to guarantee this placement.
So you see it may not be able to really even use that extra space properly and I have seen in the past customers face issues with disks on the lower capacity-nodes becoming 100% full while *technically* the vsandatastore had plenty of free space (all on the higher-capacity node of course).
A potential way of making this work would be to swap the capacity-drives in the nodes with less slots with larger drives and balance this with a higher quantity of smaller drives in the node(s) with more slots available.
thanks Bobkins , looking at a stretched cluster in this scenario, using the default FTT=1, RAID1
2 hosts currently with the pair of disk groups and 6 with single disk group.
The pair with 2 disk groups, each have a disk group that is each slightly more than double the capacity of the 6 hosts single disk group.
Wrong Doug, wrong Bob, I see what you did there
Stretched-cluster is actually likely the best-case scenario for disk-balance if going for a non-homogeneous cluster provided one of the larger capacity hosts is in each site.
If they are both on the same site then this wouldn't be of any benefit here as that added space could not be utilised as each site stores a replica of the others and the lower capacity site would be the limitation - the exception to this would be if you had a significant amount of FTT=0 data which could use that extra space on a single site.
So hopefully there is nothing preventing having one of each higher-capacity nodes at each site.