We are in to the banking sector. We have one virtual machine which is running with windows server 2008 having core banking application. As it is core banking application we don't want that virtual machine to be down. We have two ESxi host running.The possible options are like this
1) HA: If the host where my banking application running is down,the VM will restart on the next ESXi host. But here is there is a downtime of 15 Minutes.
2) FT: It is a good option, but when there is a application issue same will be replicate in the other VM also.
Is there any option to have clustering which will give the continous working of VM and my core banking application even the ESXi host failure
I think you may have answered your own question. You pointed out that FT will provide protection from a host failure, and that addresses your requirement to keep the VM running if one of your ESXi hosts dies. You're right to note that FT will simply replicate a failure within the VM, but that's an issue that you'll need to address independent from your virtualization platform.
Michael is absolutely right.
If you want to be able to roll back, you'll need to look at some other offerings. As an example, EMC RecoverPoint can do DVR-style rollback to any (reasonable) point in time for a database.