Adding nodes using switch discovery

Use switch discovery to provision nodes by specifying the switch name, port number, and network interface of the switches that are connected to the nodes that are being provisioned.

1. Configure SNMP server on the switch with administrator privilege.

   # Switch(config)# access-list 10 permit 192.168.0.20 
   # 192.168.0.20 is the IP of MN
   # Switch(config)# snmp-server group xcatadmin v3 auth write v1default 
   # xcatadmin is the group name
   # Switch(config)# snmp-server community public RO 10
   # 10 is the access-list number
   # Switch(config)# snmp-server community private RW 10 
   # 10 is the access-list number
   # Switch(config)# snmp-server enable traps
   # Switch(config)# snmp-server user xcat xcatadmin v3 auth SHA passw0rd access 10 
   # xcat is the authorized username for SNMPv3 communication, 
       "passw0rd" is sample for the password of "xcat" user

   Note: If the management node and compute nodes in the same VLAN network, configure the xcatadmin group. For example, if all nodes use VLAN ID 12, specify the following option:

   # Switch(config)# snmp-server group xcatadmin v3 auth context vlan-12

   Run the following command on management node to verify that SNMP communication is set up.

   # snmpwalk -v 3 -u xcat -a SHA -A passw0rd -l authnoPriv 192.168.0.234 
       .1.3.6.1.2.1.2.2.1.2 
   # 192.168.0.234 is the IP of switch

2. Add a switch object into switches table on the Platform Cluster Manager Community Edition management node with root privilege.

   # tabedit switches
   #switch,snmpversion,username,password,privacy,auth,linkports,sshusername,
   sshpassword,protocol,switchtype,comments,disable
   "switch","3","xcat","passw0rd",,"sha",,,,,,,

   Add the IP address of the switch to /etc/hosts:

   #echo "192.168.0.234 switch" >> /etc/hosts

3. Create a node information file /root/nodeinfo that lists all of the nodes that you want to provision. The node information file must include the NIC name, the switch name, and the switch port number.

   #node information file
   __hostname__:
   switches=eth0!switch1!1
   
   __hostname__:
   switches=eth0!switch1!2
   
   __hostname__:
   switches=eth0!switch2!3

4. Add the compute nodes to the cluster by importing a node information file in the Web Portal. You must specify a node information file that follows the formatting in the previous step. After all of the nodes are imported, the nodes are displayed in the Web Portal. For example, the node information file in step 3 creates the following nodes: switchnode00, switchnode01, and switchnode02.
   Note: When you add nodes that have a BMC or IMM network interface and supports hardware control, you must specify a network profile, which has at least two network interfaces: one Ethernet interface in the provision network and one BMC interface.
5. Start the nodes by PXE boot.