Bright Cluster – HA Setup

  1. Configure failover interface in 1st Head Node
  2. Register license for HA
  3. Prepare HA configuration using command
     $ cmha-setup
  4. Go to Setup menu to prepare configuration
  5. Check Clone Failover for clone instruction
  6. Power-on second head node, let it boot to PXE
  7. Select Rescue in PXE menu
  8. Login as root
  9. Start cloning by calling
     $ /cm/cm-clone-install --failover
  10. The second head node will reboot automatically once completed
  11. Back to primary head node.
  12. In cmha-setup menu, select Finalize.
  13. Verify HA status
     $ cmha status
  14. In cmha-setup menu, configure shared storage.


EGO issue in IBM SCF CE


[root@pcmce-co68 ~]# egosh resource list
Cannot get host info. Not logged on.

Log in to egosh shell (one time only). Default user and password is Admin.

[root@pcmce-co68 ~]# egosh user logon
user account: Admin
Logged on successfully
[root@pcmce-co68 ~]# egosh resource list
NAME status mem swp tmp ut it pg r1m r15s r15m ls
pcmce-c* ok 827M 1516M 69G 7% 258 3.1 0.2 3.3 0.7 1


Disable Serial Console Redirection in xCAT

PXE boot config:

"... console=tty0 console=ttyS0,115200 ..."

Note: serial console redirection is managed by hardware profile.

To check hardware profile:

$ tabdump nodehm

To disable: clear the entries of serialport, serialspeed and serialflow columns

$ chdef -t group -o __HardwareProfile_IBM_Flex_System_x serialport= serialspeed= serialflow=


catatan: gpfs on windows node

Here’s a general installation procedure

  1. Create AD Domain User account “root”
  2. Add root account to windows 2008 server administrators group
  3. Logon with root account
  4. Disable Windows UAC – REBOOT
  5. Disable Windows Firewall
  6. Install Role – Services for Network File System
  7. Install Feature – Subsystem for UNIX-Based Applications
  8. Download “Utilities and SDK for UNIX-based Applications_AMD64”
    Start > Program > Subsystem for UNIX-based Applications > Download Utilities for Subsystem …
    Select All Components
    Select All Unix features – SuToRoot, setuid, case sensitive
  9. Install Boostrap Installer
    run pkg-2.10.9-bootstrap60x64.exe – accept install updates
  10. Install OpenSSH
    open a C shell
    $ pkg_update -L openssh (require internet connectivity)
  11. Create ssh key and configure passwordless logon with other cluster nodes
  12. run GPFS-3.3.Windows-cz1rpen.msi – REBOOT
  13. uninstall gpfs 3.3 using “add/remove programs”
  14. install gpfs-
  15. Launch C Shell
    chown -R root /var/mmfs
    chown -R root /var/adm/log
    chown -R root /var/adm/ras
    chown -R root /usr/lpp/mmfs
  16. Activate GPFS win service,
    mmwinservctl set --account <rootaccount> --password <rootpassword>
    mmwinservctl enable
  17. On GPFS cluster manager,
    mmaddnode <win-node>
  18. Back to win node,
    mmmount <device> <driveLetter>
  19. error log is in /var/adm/ras/mmfs.log.latest

catatan: gpfs on linux node

cuma catatan installasi gpfs di linux nodes (client side)…

  1. make sure ssh connection between gpfs server and client.. kalau perlu tambahkan “StrictHostKeyChecking no” pada sebuah file “config” dalam folder “.ssh”
  2. install gpfs base rpm
  3. install gpfs update rpm
  4. building gpfs portability layer… it needs package kernel-smp-devel-<uname -r>.EL
    cd /usr/lpp/mmfs/src
    make Autoconfig
    export SHARKCLONEROOT='pwd'
    make World
    make InstallImages

    this will build some binaries in /usr/lpp/mmfs/bin
  5. at the server side, add client for gpfs service
    mmaddnode <node>
  6. at the client side, run mmstartup