The ability to scale out incrementally instead of doing a big five year capital expense purchase that hurts the budget.
In the storage world, when it's time to buy an array, you buy half that are to be populated, and you buy more disk shelves, but it's not cost effective. Or, you buy it all and don’t use some. But with VSAN, you buy x86 servers as needed and you're done, and you can scale up.
The cost. We had a directive from the CIO to check out and play with VSAN and then to do storage cheaper and better than before, which we have achieved.
The management platforms have some gaps. It's difficult to see what’s going on with the hardware at times. The only platform available doesn’t run full time, and there is a management pack but it requires a product that not many people have (vRealize Operations). So it could use more work in management areas.
Also, it lacks deduplication, so we're using a lot more storage than you necessarily need to.
Prior to deployment, make sure you check your hardware compatibility prior, along with the drivers and firmware, because any one of those three things can cause an outage.
It's very stable, but we ran into some early issues with drivers and firmware, but this is resolved now. You must be careful to be coverd in that respect, otherwise you will have issues.
We’re scaling in a very phased process, running dev test environment with just a small three node cluster, but gradually shifting.
I've never had to use it.
It was pretty straightforward, but we had some issues with drivers, although nothing in the setup. After some time, we were losing some disks that was because of driver issues.