very simple. If the controller has a low queue-depth (which I think is the case here) then your VSAN will run fine as long as there is little IO going on. But as soon as there is a lot of IO, like during maintenance, evacuation of a host etc. etc., then the controllers wil not be able to handle it, causing massive latencies, in turn causing VSAN to totally shit itself, VM's becoming unresponsive and the whole nine yards. Just like cheap consumer-grade SATA SSD's and disks tend to do. They too cannot empty their own queues fast enough and cause massive stoppages and massive latency.
My advice: don't do it. Get a couple of LSI 9702-8i HBA's. They are cheap, are supported and work very well. One note: nowadays these cards come with firmware v20 which is broken. You MUST downgrade to FW v19 to get outta trouble and be HCL compliant at the same time (FW v20 is not on the HCL for a reason...)
