Update RHCS for Debian.

debian/README.Debian

@@ -0,0 +1,31 @@
+Documentation pointers:
+
+Configuration system:
+http://fabbione.fedorapeople.org/CCS.pdf
+
+Configuration using LDAP:
+http://people.redhat.com/ccaulfie/docs/ClusterLDAP.pdf
+
+Configuration examples are located in /usr/share/doc/gfs-tools/examples/
+
+rgmanager error codes are located in /usr/share/doc/rgmanager/errors.txt
+
+------------
+
+It appears that in some hw combinations the kernel is not fast enough to handle
+cman heartbeat properly. This results in cman kicking out nodes during heavy I/O
+operations.
+
+The only known workaround is to increase the cman timeout using FOR EXAMPLE:
+
+echo $timeout_in_seconds > /proc/cluster/config/cman/deadnode_timeout
+
+The default is 21 seconds.
+
+It is important to understand that higher values of timeout will expose the cluster
+to other problems, like for example a node is not kicked out from the cluster fast
+enough and it might cause more damage than expected.
+
+NOTE that the value must never be higher than the reboot time of the node. Otherwise
+the cluster will refuse the node to rejoin the quorum and manual kick has to be done
+by the administrator.

debian/README.source

@@ -0,0 +1,6 @@
+Building cluster for Debian
+---------------------------
+
+The cluster source package uses quilt to apply and remove its patches.  Please
+refer to /usr/share/doc/quilt/README.source for information about how to use
+quilt for source packages.

debian/TODO

@@ -0,0 +1,39 @@
+Merge cman.init changes from upstream to handle xen.
+Review *.init. 
+check scsi_reserve stuff
+
+Open packaging issues
+--------------------
+
+- provide upgrade path from cluster suite 1.03 in etch:
+  * upgrade document 
+  * cman preinst: on upgrade from 1.0x, bug the user
+  * debconf template: point the user to the upgrade document, 
+    and let the user proceed at own risk
+
+- update initscripts to use lsb functions
+
+- prod upstream to write missing manpages 
+
+W: rgmanager: binary-without-manpage usr/sbin/clunfslock
+W: rgmanager: binary-without-manpage usr/sbin/rg_test
+W: gnbd-server: binary-without-manpage usr/sbin/gnbd_clusterd
+W: gnbd-server: binary-without-manpage usr/sbin/gnbd_get_uid
+W: gnbd-client: binary-without-manpage usr/sbin/gnbd_monitor
+W: gnbd-client: binary-without-manpage usr/sbin/gnbd_recvd
+W: gfs2-tools: binary-without-manpage usr/sbin/gfs2_edit
+W: gfs2-tools: binary-without-manpage usr/sbin/gfs2_umount
+W: gfs-tools: binary-without-manpage usr/sbin/gfs_debug
+W: gfs-tools: binary-without-manpage usr/sbin/gfs_edit
+W: cman: binary-without-manpage usr/sbin/fence_baytech
+W: cman: binary-without-manpage usr/sbin/fence_cpint
+W: cman: binary-without-manpage usr/sbin/fence_ibmblade
+W: cman: binary-without-manpage usr/sbin/fence_rackswitch
+W: cman: binary-without-manpage usr/sbin/fence_rps10
+W: cman: binary-without-manpage usr/sbin/fence_rsb
+W: cman: binary-without-manpage usr/sbin/fence_scsi
+W: cman: binary-without-manpage usr/sbin/fence_scsi_test
+W: cman: binary-without-manpage usr/sbin/fence_vmware
+W: cman: binary-without-manpage usr/sbin/fence_xcat
+W: cman: binary-without-manpage usr/sbin/fence_zvm
+