ESA, adaptive RAID5, stretched cluster and votes

[kastlr]

Hi wise ones,

looking for some information how ESA would handling failure scenarios in a stretched cluster environment, especially when adaptive RAID5 is in use.
In former days vSAN did use votes to figure out if a stretched vSAN Cluster could still stay up & running or would be powered off.

But with the adoption of ESA, adaptive RAID5 and different handling (under the cover) of "performance leg" vs "capacity leg" I would like to know how vSAN handles it when all new features are in place.

Regards,

Ralf

[Duncan Epping]

it still uses votes, and you can look at "perf leg" and "cap leg" as a way to get around needed a witness. Using the placement of the perf leg combined with the cap leg we can determine who owns what. I haven't done a write up of this showing the votes, but if I have some time at some point I may do this. Using "rvc" you can see the votes via the command line if you are interested.

[TheBobkin]

Just to add to what Duncan said - RVC is deprecated (and honestly I have found myself using it less and less the past few years that cannot recall using it this year) as almost everything it offered is available on any host in the cluster - in this case 'esxcli vsan debug object list' (--all if want more than the first 100 objects) shows the votes and other useful information for all objects in the cluster.