Hyper v question

[nicholab]

Hyper v layout got cluster setup at new school needs moving around this is how I think it should be laded out. Thoughts please?

 

Hyper v local

Dc fsmo Dns start first

 

Hyper v 2 local

DC Dns

 

Csv

DC with Dhcp or just DHCP or cluster on the other DC?

SIMS

SCCM

Fileserver

Print server

[mukz]

If i understand what you said correctly, sorry it was hard! Why not cluster both the hyper v servers and then virtualise all of the stuff you outlined or need. DC's can go on hyper v too, no need for physical

[nicholab]

The hyper-v server are clustered I heard it is hard getting a CSV going without a DC. Would you have DHCP server in your clustered instance? I want all of the key stuff on new servers i.e. the hyper-v hosts. I've seen discussions about putting a DC on local storage on one of your hosts and setting a rule so it will startup when the host does. I want to put my FSMO roles on the same server. Edited January 18, 2017 by nicholab

[snagrat]

DC can be clustered without issues from Server 2012R2 onwards. CSV no longer needs domain to come live

[mukz]

Ahh sorry again! Yes i have the DC's on each server locally with the DHCP failovered, was setup by my predecessor this way. Works well if i am being honest. - but when i change to 2016 cluster it was all be on the CSV

[nicholab]

Ok i need to see what type of CSV it is and that might effect the choose of placement. I also got to sort out IP addressing as the server seem to have ip addresses in each client VLAN and no routing setup on the L3 switch. Not sure what to do first IP address or hyper-V? I also have 4 DC's a bit over kill for 300 clients.

[nicholab]

Picking up this threat could you have a DC that is outside of a Cluster share volume but uses hyper-v replica? This would then mean it sits outside of the CSV but is still has better availability? Or place it on cluster storage that is not a CSV but still part of the SAN?

[Oaktech]

I have a physical DC - a HP Microserver outside the cluster on a 2 hr UPS, connected to the same switch as the cluster. That switch has a UPS on it with about 3 hours uptime, and then *everything* else is clustered in failover, also with a long run ups on it.