I want to setup a DNAC cluster, but regarding the required interfaces and ip addresses, I still do not get it completly. Tomas' ip addressing sheet helped a lot - here my summary how I understand it.
1. Cluster Port
Some RFC1918 ip addresses that are not used within my network can be used analog to server cluster heartbeat links. All nodes are in the same subnet; routing is not required since this network is isolated.I have to check the "cluster link" option for this network. Is there a cluster virtual ip required for this network?
2. Management Port
IP out of my network management network used to access the gui of the cluster - one IP for each node + VIP all in the same subnet, default gateway set if cloud link is not used, otherway static route to all my company internal ip range (I guess mostly RFC1918).
3. Cloud Port
Optional "DMZ" port to access the internet, if this cannot be done via management port - not required in my setup
4. Enterprise Port
Separte interface to connect to "the network". Since my DNAC is located in the DC, it does not have a direct connection the SDA (pilot) network (which I guess is a common setup). I'm just using one of our server subnets. Like Management, one ip per node + VIP are needed, but instead of default gw static routes to the SDA underlay network IPs are set. Can these routes be modified and added later?
Services Subnet and Cluster Services Subnet are used "inside" the DNAC - these ranges are isolated networks that just should not conflict/overlapp with the ip ranges used within the company network?.
Please let me know if I got this correct.
Solved! Go to Solution.
While watching the installation video, I got explained that management port is NOT requiered and you can use the enterprise port also for management. According to the 1.2.5 installation guide, the management port IS requiered. Does anyone know which statement is correct?
The Services Subnet & Client Services subnet needs to be a valid IP address space in a "Private" Network space. To avoid IP conflicts, this subnet must not be used elsewhere.
Private IPv4 addresses
|RFC1918 name||IP Address Range||Number of Addresses|
|24-bit block||10.0.0.0 – 10.255.255.255||16777216|
|20-bit block||172.16.0.0 – 172.31.255.255||1048576|
|16-bit block||192.168.0.0 – 192.168.255.255||65536|
I hope this helps!
This presentation is the first document marking the management port as optional :-) So the easiest setup is using the 1st 10GE as cluster link, using the 2nd 10GE as combined Manegement/Enterprise port and leave the two 1GE ports unconfigured.
Since the 10GE ports are configured as trunk and I want to use the former gui port IPs, I needed to change the vlan id on all appliances which requires a reboot of each appliance. I figured out the best way to do this is first to change the vlan id in CIMC, reboot the box and than run the maglev-config update. This sequence ansures that NTP, DNS etc. are reachable when running the update script.