#!/bin/sh

# this script needs to be installed so that statd points to it with the -H 
# command line argument. The easiest way to do that is to put something like this in 
# /etc/sysconfig/nfs:
#   STATD_HOSTNAME="myhostname -H /etc/ctdb/statd-callout"

[ -z "$CTDB_BASE" ] && {
    export CTDB_BASE="/etc/ctdb"
}

. $CTDB_BASE/functions
loadconfig ctdb
loadconfig nfs

[ -z "$STATD_SHARED_DIRECTORY" ] && {
	echo STATD_SHARED_DIRECTORY not configured. statd-callout failed.
	exit 0
}

[ -d $STATD_SHARED_DIRECTORY ] || exit 0

[ -z $NFS_HOSTNAME ] && {
	echo NFS_HOSTNAME is not configured. statd-callout failed.
	exit 0
}

case "$1" in
  add-client)
	# the callout does not tell us to which ip the client connected
	# so we must add it to all the ips that we serve
        for f in $CTDB_BASE/state/statd/ip/*; do
	    ip=`basename $f`
	    [ -d $STATD_SHARED_DIRECTORY/$ip ] || /bin/mkdir $STATD_SHARED_DIRECTORY/$ip
	    touch $STATD_SHARED_DIRECTORY/$ip/$2
	done
	;;
  del-client)
	# the callout does not tell us to which ip the client connected
	# so we must add it to all the ips that we serve
        for f in $CTDB_BASE/state/statd/ip/*; do
	    ip=`basename $f`
	    /bin/rm -f $STATD_SHARED_DIRECTORY/$ip/$2
	done
	;;
  notify)
	# we must restart the lockmanager (on all nodes) so that we get
	# a clusterwide grace period (so other clients dont take out
	# conflicting locks through other nodes before all locks have been
	# reclaimed)

	# we need these settings to make sure that no tcp connections survive
	# across a very fast failover/failback
	#echo 10 > /proc/sys/net/ipv4/tcp_fin_timeout
	#echo 0 > /proc/sys/net/ipv4/tcp_max_tw_buckets
	#echo 0 > /proc/sys/net/ipv4/tcp_max_orphans

	# rebuild the state directory for the local statd to use the correct
	# state value and to initally send notifications to all clients
	rm -f /var/lib/nfs/statd/sm/*
	rm -f /var/lib/nfs/statd/sm.bak/*
	cat $STATD_SHARED_DIRECTORY/state >/var/lib/nfs/statd/state


	# we must keep a monotonically increasing state variable for the entire
	# cluster  so state always increases when ip addresses fail from one
	# node to another
	[ ! -f $STATD_SHARED_DIRECTORY/state ] && {
		echo 1 | awk '{printf("%c%c%c%c", $0, $0/256, $0/256/256, $0/256/256/256);}' >$STATD_SHARED_DIRECTORY/state
	}
	# read current state
	STATE=`od -t d4 $STATD_SHARED_DIRECTORY/state | head -1 | sed -e "s/^[0-9]*[^0-9]*//"`
	# write current state+2 back to the state file
	# the /2 *2 are to ensure that state is odd. state must be odd.
	STATE=`expr $STATE "/" 2 "*" 2 "+" 3`
	echo $STATE | awk '{printf("%c%c%c%c", $0, $0/256, $0/256/256, $0/256/256/256);}' >$STATD_SHARED_DIRECTORY/state
	


	# we must also let some time pass between stopping and restarting the
	# lockmanager since othervise there is a window where the lockmanager
	# will respond "strangely" immediately after restarting it, which
	# causes clients to fail to reclaim the locks.
	# 
	startstop_nfslock stop > /dev/null 2>&1
	sleep 2

	# copy all monitored clients on this node to the local lockmanager
	for f in `/bin/ls $CTDB_BASE/state/statd/ip/* 2>/dev/null`; do
	    ip=`basename $f`
	    [ -d $STATD_SHARED_DIRECTORY/$ip ] && [ -x /usr/bin/smnotify ] && {
		for g in `/bin/ls $STATD_SHARED_DIRECTORY/$ip/* 2>/dev/null`; do
			client=`basename $g`
			touch /var/lib/nfs/statd/sm/$client
		done
	    }
	done

	# now start lockmanager again with the new state directory.
	startstop_nfslock start > /dev/null 2>&1

	# we now need to send out additional statd notifications to ensure
	# that clients understand that the lockmanager has restarted.
	# we have three cases:
	# 1, clients that ignore the ip address the stat notification came from
	#    and ONLY care about the 'name' in the notify packet.
	#    these clients ONLY work with lock failover IFF that name
	#    can be resolved into an ipaddress that matches the one used
	#    to mount the share.  (==linux clients)
	#    This is handled when starting lockmanager above,  but those
	#    packets are sent from the "wrong" ip address, something linux
	#    clients are ok with, buth other clients will barf at.
	# 2, Some clients only accept statd packets IFF they come from the
	#    'correct' ip address.
	# 2a,Send out the notification using the 'correct' ip address and also
	#    specify the 'correct' hostname in the statd packet.
	#    Some clients require both the correct source address and also the
	#    correct name. (these clients also ONLY work if the ip addresses
	#    used to map the share can be resolved into the name returned in
	#    the notify packet.)
	# 2b,Other clients require that the source ip address of the notify
	#    packet matches the ip address used to take out the lock.
	#    I.e. that the correct source address is used.
	#    These clients also require that the statd notify packet contains
	#    the name as the ip address used when the lock was taken out.
	#
	# Both 2a and 2b are commonly used in lockmanagers since they maximize
	# probability that the client will accept the statd notify packet and
	# not just ignore it.
        for f in `/bin/ls $CTDB_BASE/state/statd/ip/* 2>/dev/null`; do
	    ip=`basename $f`
	    [ -d $STATD_SHARED_DIRECTORY/$ip ] && [ -x /usr/bin/smnotify ] && {
		for g in `/bin/ls $STATD_SHARED_DIRECTORY/$ip/* 2>/dev/null`; do
			client=`basename $g`
#			/bin/rm -f $g
			# send out notifications from the "correct" address
			# (the same addresse as where the lock was taken out
			# on)   some clients require that the source address
			# matches where the lock was taken out.
			# also send it both as a name that the client
			# hopefully can resolve into the server ip and
			# and also by specifying the raw ip address as name.
			/usr/bin/smnotify --client=$client --ip=$ip --server=$ip --stateval=$STATE
			/usr/bin/smnotify --client=$client --ip=$ip --server=$NFS_HOSTNAME --stateval=$STATE
		done
	    }
	done
	;;
esac