xcat-core/xCAT/templates/e1350
2012-10-18 19:44:41 +00:00
..
chain.csv Have chain end in shell instead of standby 2012-08-02 12:20:08 +00:00
hosts.csv Make the e1350 templates work with nodenames like n01, not just n1 2012-07-31 18:36:50 +00:00
ipmi.csv -Update sample files with a more resiliant regex for node number extraction 2008-05-14 19:53:15 +00:00
mp.csv -Update sample files with a more resiliant regex for node number extraction 2008-05-14 19:53:15 +00:00
nodehm.csv
nodepos.csv -Remove '-' from nodepos.u of idataplex sample entry 2008-05-14 20:04:09 +00:00
noderes.csv -Change default pxe handler to xnba 2009-09-28 13:04:52 +00:00
nodetype.csv Type commit message here…updating the e1350 templates 2012-10-18 19:44:41 +00:00
passwd.csv
README Make the e1350 templates work with nodenames like n01, not just n1 2012-07-31 18:36:50 +00:00
servicenode.csv Type commit message here…updating the e1350 templates 2012-10-18 19:44:41 +00:00
switch.csv -Update sample files with a more resiliant regex for node number extraction 2008-05-14 19:53:15 +00:00

README for: /usr/share/xcat/templates/e1350 directory
-----------------------------------------------------

The set of *.csv files in this directory are sample xCAT table values that
you may wish to use as templates for defining your xCAT cluster.
The data is based on defining a cluster using the address scheme given in:
http://download.boulder.ibm.com/ibmdl/pub/systems/support/system_x_pdf/intelligent_cluster_factory_settings_05.24.12b.pdf.
Additional information on using xCAT 2.0 can be found at 
http://sourceforge.net/apps/mediawiki/xcat/index.php?title=XCAT_iDataPlex_Cluster_Quick_Start.

Check the 1350 default templates to see if they apply to your environment.
If you decide to use some of these templates, modify them, or create your own,
you can first edit the files and then load them into the database, or load the
files first and then edit the database once they are loaded.

To load a file, use the tabrestore command on each template to load the tables 
in the database.

To load the database tables:
   for i in *.csv; do tabrestore $i; done

To verify that a table has been correctly loaded, run "tabdump <tablename>".   
For example: 
   tabdump nodetype 


You can add nodes to the database by using the nodeadd command. 
See the example below and details in http://sourceforge.net/apps/mediawiki/xcat/index.php?title=XCAT_iDataPlex_Cluster_Quick_Start.  

For example, to add 2048 IPMI rackmount nodes which happen to start 
at the first U of the first rack, with each rack having a 1U switch
and 41 servers plugged into that switch, numbered 1-41 in order,
with the nodes named node1-node2048, and members of groups named
41nodeperrack,41perswitch,ipmi,compute,all, run the following:

  nodeadd node1-node2048 groups=41nodeperrack,41perswitch,ipmi,compute,all

After that, other tables will setup according to that scheme:
nodels node1235 nodepos.u nodepos.rack switch.switch switch.port ipmi.bmc hosts.ip
node1235: hosts.ip: 172.20.131.5
node1235: nodepos.u: 5
node1235: nodepos.rack: 31
node1235: switch.switch: switch31
node1235: switch.port: 5
node1235: ipmi.bmc: bmc1235

Note: the hosts table is only used if you use the makehosts command to setup /etc/hosts.

An example with blades, 56 to a rack and padding out to four digits:
nodeadd node0001-node2048 groups=blade,compute,all,56nodeperrack


If wanting to use makehosts for non-server equipment, nodeadd can help for that as well with these templates:
(Supplementing the previous rackmount example)

nodeadd switch1-switch50 groups=switch
nodeadd bmc1-bmc2048 groups=41bmcperrack
makehosts switch,41bmcperrack

(And to make hosts file entries for 147 AMMs for the 2048 blades above)
nodeadd amm1-amm147 groups=amm
makehosts amm