How to avoid a "race condition" after crashing a server rack with Hyper-V over SMB 3.0 deployed?

Question

Active Directory

I came across a very unpleasant feature of Hyper-V running over SMB 3.0 on a small scale on a single object.
There is a rack with equipment: a UPS, several Hyper-V hosts, a large noisy file server with many disks that acts as storage for Hyper-V. For the past couple of weeks, this facility has been regularly losing power for longer periods of time than the UPS can pull out, so automatic shutdown of all equipment is configured at a critical battery level.
Other than the file server and Hyper-V hosts, there are no physical servers. The domain controllers are virtualized, one on the local drives of each Hyper-V host.
When turning on all the hardware after the "emergency shutdown", something like a race condition begins: if the Hyper-V hosts booted before the file server, all virtual machines go into an emergency state due to loss of storage; if the file server started before the Hyper-V hosts, it will not detect domain controllers on the network and will deny Hyper-V hosts access to the file share, after which the virtual machines, again, go into an emergency state.
Electricity is most often cut down at night, so the morning at the IT department starts much earlier than we would like.
What are the possible solutions to the problem?

Answer 1

so no one answered for so many years? :)
If you don't bother, then the solutions are as follows:
1 - for all hypervisors in the BIOS, enable a full test of RAM. this will give them a fun activity for a while.
or
2 - in the parameters of the guests, set the desired start delay.