This document explains how to quickly install openQRM on a RedHat/FedoraCore/Whitebox/Centos linux server.

1. openQRM will run on a single server, or two openQRM servers as a HA cluster via the HA plugin.  The system will provision your choice of Linux distros across your set of PXE boot enabled nodes.    openQRM will provide tftp and dhcp services on your management lan and manage the kernels and filesystem images on your nodes via a web GUI.  All machines must to be on the same flat network which we call the management network.  For ease of documentation we will assume eth0 is the management nic and eth1 provides that node's services to external users.  We will build the openQRM server and use it to deploy kernel and filesystem images from the openQRM server to PXE boot enabled nodes on the management network.  All the steps below assume you have a root prompt.
   
   In our examples below, the ip address of eth0 on the openQRM server is 192.168.1.200.  The DHCP server on the openQRM server will issues IP addresses in the range 192.168.1.50-80.
   
   
2. Please ensure that http/apache, dhcpd and tftp servers are not running on your openQRM server prior to openQRM installation.  openQRM will provide these services.
   
   
3. Install a MySQL v5.0.22 server on the openQRM machine.  If you have an existing one in your network then you can use it instead.  Detailed instructions can be found here.
   
   
4. In this example we will use a local NFS server for storing the filesystem images which will deployed to your nodes.  On the openQRM server, export the root partition by adding:
   "/  *(rw,no_root_squash,insecure,async)" to your /etc/exports file and start NFS with "service NFS restart".  If you have an existing NFS server on your network, you can use it instead.
   
   
5. Download the openQRM PRM package and the DHCP and TFTP plugins.  As of this writing, you would get the following three files:  qrm-core-base-2.2-3.i386.rpm, qrm-plugin-dhcpd-2.2-3.i386.rpm, qrm-plugin-tftpd-2.2-3.i386.rpm.
   
   
6. Install the rpms.  "rpm -ivh qrm*.rpm"
   
   
7. In the /opt/qrm directory you will find the qrm-installer program.  Run this to create the qrm.xml file which contains your site specific details.  You will need to perform three steps with this ncurses-based utility.  You will navigate the utility with your tab, arrow and space keys.
   
   
   1. Basic Configuration.  Press enter here and answer the questions asked.  The first is which NIC to use.  In our example here, the answer is 'eth0'.  The default provided is 'eth0:qrm', so you will need to change it.  for all other questions the default provided is acceptable.
      
      
   2. Plugn selection.  Using your spacbar, put an 'X' next to the DHCP and TFTP plugins.  Choose OK.
      
      
   3. Plugin configuration.  Here you see two plugins.
      
      
      1. Choose DHCPD.  You will want to use values specific to your network here.  In our example, this dhcpd server provides its services to the 192.168.1.0/24 network.
         
         
         • Network address: 192.168.1.0
         • Netmask: 255.255.255.0 
           
         • Network Broadcast address: 192.168.1.255
         • Gateway: 192.168.1.1
         • DNS: 192.168.1.1
         • QRM Server address = the address of this server. 
         • TFTPD address = same as above
         • Domain name for dhcp clients = arbritary
         • dhcp range start address = 192.168.1.50
         • dhcp range end address = 192.168.1.80
         • Choose OK
           
           
      2. the TFTPD plugin requires no configuration.
         
         
      3. Choose OK and exit.
         
      
      
8. After that the openQRM installation utility "/opt/qrm/qrm-installer" will start the installation procedure. It  will be asked the following questions:
   
   
   • Enter License number - press <enter>
   • Enter name - press <enter>
   • review the General and Database parameters and if correct press 'Y'.
   • review the DHCP parameters and if correct press 'y'
   • download the manual by pressing 'Y'
     
   
   When the installation is finished, the openQRM-server is running and ready to use.  Point your browser to localhost.  The initial login information is:
   
   
• Username: qrm
• Password: qrm
  
  
9. Now we boot up the first node in the openQRM-network.  Simply make sure the bios is set to do "PXE-boot" and power it on. 
   
   
   • Power on the node
   • The node sends a dhcpd/pxe request
   • The node receives an ip address and boot-file from the dhcpd-server
   • The node boot the openQRM 'idle' kernel + initrd and becomes available in the openQRM GUI.
     
     
10. Now we will create a custom kernel from OS running on the openQRM server.  We will use this kernel-image to deploy onto your managed nodes.  We create this kernel-image from the command line or from the GUI interface.
    
    
    1. /opt/qrm/bin/qrm-boot-image create --kernel-version `uname -r` --boot-image my-kernel --physical-name mykernel
    2. The running kernel from the openQRM server gets processed and saved into the openQRM system.
    3. Using the GUI, you can examine the new entry in the Kernels section under Management Tools-->Kernels
       
       
11. Now we create a storage server which will be used to save a filesystem-image onto.  In this example, the storage server will point to the NFS server running on the openQRM server.  This can be done with the GUI or on the command line. 
    
    
    • /opt/qrm/bin/qrm-cli storage add --name openqrmNFS --type NFS --ip <ip-address-of-openQRM-server>
      
    
    
12. Next, we create a custom filestem-image from the root filesystem funning on the openQRM server.   This must be done from the command line.
    
    
    •  /opt/qrm/bin/qrm-filesystem-image create -u qrm -p qrm -s myfsimage -i 192.168.1.200 -t 192.168.1.200:/image3 --storage-server openqrmNFS -l / -e /tmp
      
      
13. Now we create a 'Virtual Environment'.  A VE is a paring of a kernel image with a filesystem image.  When you start a VE, openQRM will deploy your kernel and filesystem image onto your node.  You can create a VE using the GUI or from the command line.
    
    
    • To configure and start a new virtual-environment, access the openQRM-server GUI. Choose "virtual environments" -> "new virtual environment". Fill out and save the configuration form; choose the previous created boot- and file system-image.
      
      
14. Power up one or more nodes which are configured to PXE boot from the network card and watch them boot up the openQRM 'idle' kernel.  These nodes are visable as idle resources.
    
    
15. You can now start the just created new virtual environment using the GUI.  Once started, you will see one of your nodes cleanly reboot the idle kernel image and boot up using the kernel and filesystem images you specified when you created the VE.  Once the machine is booted up, you can ssh into the box using the ip address detailed in the GUI. 

Comments (0)

Subscribe to this comment's feed