Typically, in order to get a chassis into a particular subcluster, you boot the node via the CSN as normal, and use the configuration page (example: http://[CSN IP]:8090/services/cs/configuration) to put the booted node into the desired subcluster. You will then have to reboot the chassis.
You might notice that this requires two bootup sequences and this might not be desirable. It is possible that your cluster may rebalance streams to the default (undesirable) subcluster during this time period (the time it takes to configure the subcluster and then start a reboot).
To prevent this temporary condition, you can configure the CSN to boot the nodes into their proper subcluster on first boot.
When you manually configure subclusters on a CSN, two types of files are created in /var/opt/caringo/netboot/content/nodeconfigs/:
2. sn[mac address].cfg files- 1 for each NIC on each chassis
The subclusters.cfg file looks like:
>cat subclusters.cfg West East
It includes only a list of available subclusters and you do not have to modify this.
The other files look like this, where the MAC address of the Swarm node's NIC is 00:0c:29:24:22:d5
>cat sn000c292422d5.cfg ### # Unlabeled # Parameter set written at 18-Jul-16 15:18 # [ ] archiveMode = 0 subcluster = East
The only parameter we really need here is the subcluster parameter.
In order to put a new chassis into a particular subcluster, create a new file called: sn[mac address].cfg where [mac address] is the MAC address of the NIC on the chassis without any spaces, colons, dashes, etc. If you have multiple NICs connected, you need to make a file for each MAC address.
Example, if you have a new chassis with a NIC with the MAC address of 01:12:23:34:45:56 that you wanted to put in the East subcluster, you could type: