GFS bug fix update

Updated GFS packages that resolve module loading issues and other various
bugs are now available for Red Hat Enterprise Linux 3 Update 5, kernel
release (2.4.21-32.EL).

The Red Hat Global File System (GFS) provides high availability of critical
server applications in the event of planned or unplanned system downtime.

These updated GFS packages are in sync with the latest kernel
(2.4.21-32.EL). The modules in earlier GFS packages fail to load because
they do not match the running kernel. It is possible to force load the
modules; however, this update alleviates the need to do so.

In addition, the following bugs have been discovered and addressed since
the previous release:

- Fixed a problem in which lock_gulmd logged out of the master on a client
node while that client had GFS mounted, causing the cluster to hang.
- Fixed a problem in which gulm_tool getstats run on a client reported the
existence of a master if the client was logged into the master when the
master node dropped into arbitrating. Newly joined clients would then be
unable to see the Master.
- Allowed lock_gulm.o is to change its verbosity level.
- Fixed luck_gulmd to allow the user to specify what interface it would
listen on.
- Fixed a race condition in ccsd in which a child process died but the
parent did not recognize it, which resulted in a zombie child process and a
parent process that never stopped when trying to daemonize.
- Fixed a problem in which a gulm lock server ran on GFS clients after the
master server died.
- Added init scripts for gnbd to ease startup and configuration
- Fixed a problem which could cause gfs_mkfs to be confused by certain
file system sizes.
- Fixed a problem in which the lt_high_locks setting in cluster.ccs was
ignored.
- Fixed a problem in which pool attempted to do multipath on powerpath
devices, which do their own multipathing

Users are advised to upgrade to these latest GFS packages, which are
updated for use with the 2.4.21-32.EL kernel.

Before applying this update, make sure all previously released errata
relevant to your system have been applied.

To update all RPMs for your particular architecture, run:

rpm -Fvh [filenames]

where [filenames] is a list of the RPMs you wish to upgrade. Only those
RPMs which are currently installed will be updated. Those RPMs which are
not installed but included in the list will not be updated. Note that you
can also use wildcards (*.rpm) if your current directory *only* contains the
desired RPMs.

Please note that this update is also available via Red Hat Network. Many
people find this an easier way to apply updates. To use Red Hat Network,
launch the Red Hat Update Agent with the following command:

up2date

This will start an interactive process that will result in the appropriate
RPMs being upgraded on your system.

129193 - cluster can hang if lock_gulmd logs out on mounted client
129879 - gulm_tool on client reports Master node when there is no master
130358 - lock_gulm.o verbosity is not configurable
131142 - unable to select which interface to bind to from lock_gulmd
145121 - ccsd can get stuck on startup with zombie child
148029 - running gulm lockserver on mounters with master failure can mixup sockets
150029 - Pool does not see all powerpath devices
150365 - gnbd init.d scripts needed
153321 - node locks can deadlock cluster
154902 - gfs_mkfs can get confused by certain filesystem sizes
156872 - lt_high_locks setting

GFS, package