Red Hat GFS 6.1 Release Notes

Copyright(c) 2005 Red Hat, Inc.
	-------------------------------------------------------

November 3, 2005

Introduction

   The following topics are covered in this document:
   
     o Changes to GFS 6.1
     
     o Important Notes

     o Related Documentation
     
   
Changes to GFS 6.1

   This section describes changes for GFS 6.1 and includes the following
   information:

     o Cluster infrastructure provided by Red Hat Cluster Suite
     
     o Red Hat Cluster Suite lock architectures
     
     o New volume manager, LVM2

     o Enhanced gfs_fsck performance and changes to the gfs_fsck command
     
     o Withdraw individual mount points
     
     o Increased total storage supported


   Cluster infrastructure provided by Red Hat Cluster Suite

    This release of Red Hat GFS uses the cluster infrastructure of Red Hat
    Cluster Suite, taking advantage of the Red Hat Cluster Suite configuration
    file and cluster graphical user interface (GUI), system-config-cluster. For
    information about configuring and managing Red Hat Cluster Suite, refer to
    Red Hat Cluster Suite Configuring and Managing a Cluster. Previous versions
    of Red Hat GFS provided a cluster infrastructure that was exclusive to GFS
    (even though used with Red Hat Cluster Suite). Additionally, configuration
    files in earlier versions of Red Hat GFS were created and maintained via
    text editors only (that is, no GUI was available).
    

   Red Hat Cluster Suite lock architectures

    Via Red Hat Cluster Suite, GFS can use the following lock architectures:

     o DLM (Distributed Lock Manager), new for Red Hat GFS 6.1 -- DLM provides
       lock management throughout a Red Hat cluster, requiring no nodes to be
       configured as lock management nodes (contrasted to GULM, which does
       require certain nodes to be configured as lock management nodes).
  
     o GULM (Grand Unified Lock Manager) -- A client/server lock architecture
       that is compatible with Red Hat GFS 6.0.
  
     o Nolock -- For single node operation only.
    

   New volume manager, LVM2
    
    The pool volume manager in earlier releases of Red Hat GFS is replaced with
    LVM2 for this release. For information about converting pool volumes to the
    LVM2 format and upgrading from GFS 6.0 to GFS 6.1, refer to "Appendix A
    Upgrading GFS" of the Red Hat GFS 6.1 Administrator's Guide.


   Enhanced gfs_fsck performance and changes to the gfs_fsck command

    The gfs_fsck function performs 10 times as fast as gfs_fsck in earlier GFS
    releases. (This enhancement has been included in a recent update to Red Hat
    GFS 6.0, also.) In addition, the enhanced gfs_fsck function includes changes
    to certain command options. The gfs_fsck command has changed in the
    following ways:
                                                                                
     o You can no longer set the interactive mode with Ctrl+C. Pressing
       Ctrl+C now cancels the gfs_fsck command. Do *not* press Ctrl+C unless
       you want to cancel the command.
                                                                                
     o You can increase the level of verbosity by using the -v flag. Adding a
       second -v flag increases the level again.
                                                                                
     o You can decrease the level of verbosity by using the -q flag. Adding a
       second -q flag decreases the level again.
                                                                                
     o The -n option opens a file system as read-only and answers 'no' to any
       queries automatically. It provides a way to try the command to reveal
       errors without actually trying to change anything.
                                                                                
    Refer to the Red Hat GFS 6.1 Administrator's Guide, "4.12. Repairing a File
    System" and the gfs_fsck man page, gfs_fsck(8), for additional information.


   Withdraw individual mount points
   
    This release allows individual GFS mount points to gracefully discontinue
    operations on a node without causing that node to panic. This feature
    provides the ability to continue operations with unaffected file systems on
    that node. The feature can be overridden to allow a node to panic, thereby
    providing more information for troubleshooting. For more information, refer
    to the mount command option, oopses_ok, in the Red Hat GFS 6.1
    Administrator's Guide, "Table 4-2".


   Increased storage supported

    Red Hat GFS now supports 8 terabytes of storage per GFS file system.
    For more information about Red Hat GFS requirements, refer to the
    Red Hat GFS 6.1 Administrator's Guide, "Chapter 2. System Requirements".


Important Notes    
   
   For information about upgrading from GFS 6.0 to GFS 6.1, refer
   to "Appendix A Upgrading GFS" of the Red Hat GFS 6.1 Administrator's Guide.
   
   Multipath GNBD is not available with this releases of Red Hat GFS 6.1.
   That is, device mapper multipath (dm-multipath) cannot use GNBD. GNBD without
   multipath *is* available.


Related Documentation
   
   The following related documentation is available at 
   http://www.redhat.com/docs/manuals/csgfs/:
   
   o Red Hat GFS 6.1 Administrator's Guide

   o Red Hat Cluster Suite Configuring and Managing a Cluster
   
   o Release notes for Red Hat Cluster Suite 4