Appendix B. Upgrading GFS

This appendix contains instructions for upgrading GFS 5.2.1 to GFS 6.0 software.


If you are using GFS with Red Hat Cluster, the order in which you upgrade GFS compared to other Red Hat Cluster installation and configuration tasks may vary. For information about installing and using GFS with Red Hat Cluster Suite, refer to Appendix A Using Red Hat GFS with Red Hat Cluster Suite.

To upgrade the software follow these steps:

  1. Halt the cluster nodes and the lock servers. The remaining steps require that the GFS cluster be stopped (all GFS nodes shut down). Stopping the GFS cluster consists of the following actions:

    1. Unmount GFS file systems from all nodes.

    2. Stop lock servers.

    3. Stop ccsd on all nodes.

    4. Deactivate pools.

    5. Unload kernel modules.

  2. Install new software. This step consists of the following actions:

    Reference: Chapter 3 Installing GFS

    1. Install (or verify that) the Red Hat Enterprise Linux 3 Update 2 kernel (the stock "2.4.21-15.EL" kernel) is installed.

    2. Install perl-Net-Telnet RPM.

    3. Install GFS 6.0 RPMs.

  3. Load new kernel modules on GFS nodes.

    Reference: Chapter 3 Installing GFS


    insmod pool 
    insmod lock_harness 
    insmod lock_gulm 
    insmod gfs
  4. (Optional) Modify CCS files.

    With the cluster being shut down, if you need to make changes to the Cluster Configuration System (CCS) files, you have the option of doing that now. In addition, you can remove the licens.ccs file from the CCA — GFS 6.0 requires no license.ccs file.


    Although GFS 6.0 requires no license.ccs file, you can safely leave the license file in the CCA.


    You can use the ccs_tool extract command to extract the Cluster Configuration System (CCS) files for modification.

  5. (Optional) Activate pools on all nodes.

    Command usage: pool_assemble -a

    Reference: Section 5.6 Activating/Deactivating a Pool Volume


    pool_assemble -a
  6. (Optional) Create CCS archive on CCA device. The CCS archive is created from the directory of new CCS files as described in Step 5.

    Command usage: ccs_tool create Directory Device

    Reference: Section 7.1 Creating a CCS Archive


    ccs_tool create /root/alpha/ /dev/pool/alpha_cca
  7. Start ccsd on all nodes.

    This includes all GFS nodes and all nodes that will run the LOCK_GULM server.

    Command usage: ccsd -d Device

    Reference: Section 7.2 Starting CCS in the Cluster


    ccsd -d /dev/pool/alpha_cca
  8. Start LOCK_GULM server.

    Start lock_gulmd on all nodes.

    Command usage: lock_gulmd

    Reference: Section 8.2.3 Starting LOCK_GULM Servers


  9. Mount GFS file systems on all GFS nodes.

    Command usage: mount -t gfs BlockDevice MountPoint

    Reference: Section 9.2 Mounting a File System


    mount -t gfs /dev/pool/pool0 /gfs