Replica - 2012 - configure to not start on reboot

Question

I want to configure Replica in 2012 R2 so that if there is a reboot - planned or unplanned - Replica does not start up automatically - there has to be manual intervention. So for instance if Replica source server goes down as a result of a power out situation
when it comes back up it will not immediately be replicating to the target but I will have to re-initiate it. 
If that's not possible then at least on reboot I want there to be a delay  in Replica being reinitiated.
If the above can't be done as a choice within Replica then I would like to be able to
1) make the services that Replica relies on to work 'manual' so they are not automatically restarted on reboot
2) if 1 is not possible then I want to make it so that Hyper-V itself does not start on reboot - again if necessary by making the services it depends on - I believe the main one being the Hyper-V Virtual Machine Management Service.

Answer 1

Hi Arlester,
Let me answer this in a few stages. The short answer to your question is:  there is no out-of-the-box way to do this today. Any solution to solve your problem will involve some amount of scripting.
1) Regarding file corruption:
The way Hyper-V Replica works is that if there is a possible VHD corruption detected, we set the VMs to resynchronize. So if your VM starts replicating on host restart, it means that the VHD is okay. Of course, that says nothing about the application using
the VHD... and it is quite possible that there are inconsistencies at an application-level that would still be consistent at the VHD level.
A good practice here would be to schedule the resynchronization or make it manual so that it is not triggered immediately. The settings for resynchronization can be found under VM Settings --> Replication --> Resynchronization.
2) Pausing replication:
In the scenario that you have described, you seem to be looking for a way to pause the replication on a host restart. There are two places where you can pause the replication: on the source or on the destination. The suggestion I would give is to write a script
to monitor your source and pause VMs on the destination when the source is unreachable. Pausing replication on the destination will reject incoming packets from the source, while pausing replication on the source will stop sending the packets itself. The reason
I think pausing on the destination is better is because you will need a VM object to trigger this action... and on the source this needs to be done as soon as the service is up. This can lead to timing issues which can be easily avoided by pausing on the destination
site.
You can use PowerShell scripts or create a SCO runbook to monitor and pause the VMs.
Hope this helps.