Before I get into the question, let me explain what exactly I am trying to do. <o:p></o:p>
I have a two-node failover cluster (Windows Server 2016) on VMware and I am setting up SQL Server Always on AG with Primary and Secondary replica on it. I don't want to use shared disk architecture since it will break my vMotion capabilities and Always on AG doesn't require any shared storage. Now I have to configure DTC for my SQL instances and question here is do we actually need to use a "Clustered DTC" with Always on AG? I am using SQL 2017 and I see we have per database DTC support but I am confused enabling this option and leaving it to the local DTC, in the event of a failover to the secondary what happens to any in doubt transactions.
As per this document the database will go into suspect mode and I don't want that.
Now the question is, do we need clustered DTC in this scenario (My applications for sure are doing distributed transactions) and if yes, what about the storage for DTC? I would again need to add a shared disk to be used by the DTC which renders the entire solution invalid. I also see the option of using S2D pool, but as per the VMware docs, vMotion is not tested with S2D and I am not sure whether to go with it or not.<o:p></o:p>
Kindly advise. <o:p></o:p>