Failover Setup for DNS, DHCP, Storage (Files)
Hi. We have currently 2 Servers (Dell PowerEdge R530) and a Windows Server 2016 Standard OS. We want a high availability setup. If one server breaks down, then the other one will take over directly and the shared files, DNS, DHCP are still accessible. We want to use 2 physical servers only and get rid of ISCSI configured storage server. This could be possible with "storage spaces direct" but the required OS is Windows Server Data Center edition. What will be the approach for this setup? Thank you in advance.
you can create additional domain controller with DHCP failover configuration and paralelly you can create storage server using Hyper-V ... you can create clustered Hyper-V storage server
for DHCP failover see the following link
https://www.itprotoday.com/windows-8/configuring-dhcp-failover-windows-server-2016
for failover clusetering
http://www.msserverpro.com/implementing-failover-clustering-windows-server-2016-hyper-v/
all the best
for DHCP failover see the following link
https://www.itprotoday.com/windows-8/configuring-dhcp-failover-windows-server-2016
for failover clusetering
http://www.msserverpro.com/implementing-failover-clustering-windows-server-2016-hyper-v/
all the best
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Agree with andyalder, dfs duplicates the storage use each will gave a copy of the data. Note backup are important dfs is not a backup solution.
An errand file deletion on one, will quickly propagate to the other.....
Redundancy is built-in in Ad DC, and DNS ..
The DHCP redundancy can be achieved through split scope with assigning/introducing a delay (properties of DHCP server, part of conflict detection/resolution)
If you have multiple scopes, segments, both servers have to be duplicates using excluded ranges opposite on each.
You included information, but not how it is being used, so it is left to each responder to guestimate .....
This does not exclude the implementation using VMs, as Kevin and sajid mentioned.
Much depends on whether the 530's specs CPU, memory, and storage meet the needs of the ... And could sustain loads/performance for the duration prior to the upgrade cycle.
An errand file deletion on one, will quickly propagate to the other.....
Redundancy is built-in in Ad DC, and DNS ..
The DHCP redundancy can be achieved through split scope with assigning/introducing a delay (properties of DHCP server, part of conflict detection/resolution)
If you have multiple scopes, segments, both servers have to be duplicates using excluded ranges opposite on each.
You included information, but not how it is being used, so it is left to each responder to guestimate .....
This does not exclude the implementation using VMs, as Kevin and sajid mentioned.
Much depends on whether the 530's specs CPU, memory, and storage meet the needs of the ... And could sustain loads/performance for the duration prior to the upgrade cycle.
ASKER
Hi experts. Thank you for your insights. I think I will consider using the DFS.
How much monitoring of the system are you able to do?
Does it really have to be automatic?
First and foremost, you need to have a good data protection and recovery plan in place. For example, maybe that means that you run everything as a VM, and back the VM(s) from the first server to the second server. Backups should also be sent offsite on a regular basis.
VMs can be replicated from one host to another. In event of host failure, the VM(s) can be pretty quickly restarted on the second host. This is not a substitute for backups.
You can have VMs setup in a high availability cluster. In event of host failure, the VM still crashes and will be unavailable for a minute or so until it reboots. Again, you still need backups. This also requires shared storage. The shared storage can be actual physical shared storage such as iSCSI, NFS, or SMB, or virtual shared storage such as what is offered by Starwind Software and VMware VSAN. You still need backups.
There is Windows failover clustering at the file server level, which has a faster time to recovery and automatic failover, but again requires shared storage. It would normally be done with VMs these days. Again, you should have backups first.
Setting up anything other than backup and recovery is non-trivial, and is historically expensive. If you can't justify spending an additional $10,000, you probably should be sticking with just backup and recovery. Solutions such as Veeam can get you back up and running very quickly.