#!/bin/sh
#
# ocf:pacemaker:ping resource agent
#
# Copyright 2009-2023 the Pacemaker project contributors
#
# The version control history for this file may have further details.
#
# This source code is licensed under the GNU General Public License version 2
# or later (GPLv2+) WITHOUT ANY WARRANTY.
#

#######################################################################
# Initialization:

: ${OCF_FUNCTIONS:="${OCF_ROOT}/resource.d/heartbeat/.ocf-shellfuncs"}
. "${OCF_FUNCTIONS}"
: ${__OCF_ACTION:="$1"}

# Explicitly list all environment variables used, to make static analysis happy
: ${OCF_RESKEY_CRM_meta_timeout:="20000"}
: ${OCF_RESKEY_CRM_meta_globally_unique:="false"}
: ${OCF_RESKEY_name:="pingd"}
: ${OCF_RESKEY_dampen:="5s"}
: ${OCF_RESKEY_attempts:="3"}
: ${OCF_RESKEY_multiplier:="1"}
: ${OCF_RESKEY_debug:="0"}
: ${OCF_RESKEY_failure_score:="0"}
: ${OCF_RESKEY_use_fping:="1"}
: ${OCF_RESKEY_host_list:=""}
: ${OCF_RESKEY_options:=""}
: ${OCF_RESKEY_timeout:=""}

#######################################################################

meta_data() {
     cat <<END
<?xml version="1.0"?>
<resource-agent name="ping" version="@VERSION@">
<version>1.1</version>

<longdesc lang="en">
Every time the monitor action is run, this resource agent records (in the CIB) the current number of nodes the host can connect to using the system fping (preferred) or ping tool.
</longdesc>
<shortdesc lang="en">node connectivity</shortdesc>

<parameters>

<parameter name="pidfile" unique-group="pidfile">
<longdesc lang="en">PID file</longdesc>
<shortdesc lang="en">PID file</shortdesc>
<content type="string" default="${HA_VARRUN%%/}/ping-${OCF_RESOURCE_INSTANCE}" />
</parameter>

<parameter name="dampen" reloadable="1">
<longdesc lang="en">
The time to wait (dampening) further changes occur
</longdesc>
<shortdesc lang="en">Dampening interval</shortdesc>
<content type="integer" default="5s"/>
</parameter>

<parameter name="name" unique-group="name">
<longdesc lang="en">
The name of the attributes to set.  This is the name to be used in the constraints.
</longdesc>
<shortdesc lang="en">Attribute name</shortdesc>
<content type="string" default="pingd"/>
</parameter>

<parameter name="multiplier">
<longdesc lang="en">
The number by which to multiply the number of connected ping nodes by
</longdesc>
<shortdesc lang="en">Value multiplier</shortdesc>
<content type="integer" default="1"/>
</parameter>

<parameter name="host_list" required="1">
<longdesc lang="en">
A space separated list of ping nodes to count.
</longdesc>
<shortdesc lang="en">Host list</shortdesc>
<content type="string" default=""/>
</parameter>

<parameter name="attempts" reloadable="1">
<longdesc lang="en">
Number of ping attempts, per host, before declaring it dead
</longdesc>
<shortdesc lang="en">no. of ping attempts</shortdesc>
<content type="integer" default="3"/>
</parameter>

<parameter name="timeout" reloadable="1">
<longdesc lang="en">
How long, in seconds, to wait before declaring a ping lost
</longdesc>
<shortdesc lang="en">ping timeout in seconds</shortdesc>
<content type="integer" default="2"/>
</parameter>

<parameter name="options">
<longdesc lang="en">
A catch all for any other options that need to be passed to ping.
</longdesc>
<shortdesc lang="en">Extra Options</shortdesc>
<content type="string" default=""/>
</parameter>

<parameter name="failure_score">
<longdesc lang="en">
Resource is failed if the score is less than failure_score.
Default never fails.
</longdesc>
<shortdesc lang="en">failure_score</shortdesc>
<content type="integer" default=""/>
</parameter>

<parameter name="use_fping">
<longdesc lang="en">
Use fping rather than ping, if found. If set to 0, fping
will not be used even if present.
</longdesc>
<shortdesc lang="en">Use fping if available</shortdesc>
<content type="boolean" default="1"/>
</parameter>

<parameter name="debug" reloadable="1">
<longdesc lang="en">
Enables to use default attrd_updater verbose logging on every call.
</longdesc>
<shortdesc lang="en">Verbose logging</shortdesc>
<content type="string" default="false"/>
</parameter>

</parameters>

<actions>
<action name="start"   timeout="60s" />
<action name="stop"    timeout="20s" />
<action name="monitor" depth="0"  timeout="60s" interval="10s"/>
<action name="meta-data"  timeout="5s" />
<action name="validate-all"  timeout="30s" />
<action name="reload-agent"  timeout="20s" />
</actions>
</resource-agent>
END
}

#######################################################################

ping_conditional_log() {
    level="$1"; shift
    if [ $OCF_RESKEY_debug -gt 0 ]; then
        ocf_log "$level" "$*"
    fi
}

ping_usage() {
    cat <<END
usage: $0 {start|stop|monitor|validate-all|reload-agent|meta-data}

Expects to have a fully populated OCF RA-compliant environment set.
END
}

ping_start() {
    ping_monitor
    if [ $? -eq $OCF_SUCCESS ]; then
        return $OCF_SUCCESS
    fi
    touch "${OCF_RESKEY_pidfile}"
    ping_update
}

ping_stop() {

    rm -f "${OCF_RESKEY_pidfile}"

    attrd_updater -D -n "$OCF_RESKEY_name"

    return $OCF_SUCCESS
}

ping_monitor() {
    if [ -f "${OCF_RESKEY_pidfile}" ]; then
        ping_update
        if [ $? -eq 0 ]; then
            return $OCF_SUCCESS
        fi
        return $OCF_ERR_GENERIC
    fi
    return $OCF_NOT_RUNNING
}

ping_validate() {
    # Host-specific checks
    if [ "$OCF_CHECK_LEVEL" = "10" ]; then
        
        # Is the state directory writable?
        state_dir=$(dirname "$OCF_RESKEY_pidfile")
        touch "$state_dir/$$"
        if [ $? -ne 0 ]; then
            ocf_log err "Invalid location for 'state': $state_dir is not writable"
            return $OCF_ERR_ARGS
        fi
        rm "$state_dir/$$"

        # Does the ping binary exist?
        check_binary ping

    fi

# Pidfile better be an absolute path
    case "$OCF_RESKEY_pidfile" in
        /*) ;;
        *) ocf_log warn "You should use an absolute path for pidfile not: $OCF_RESKEY_pidfile" ;;
    esac

# Check the host list
    if [ -z "$OCF_RESKEY_host_list" ]; then
        ocf_log err "Empty host_list.  Please specify some nodes to ping"
        exit $OCF_ERR_CONFIGURED
    fi

    # For fping allow only same IP versions or hostnames
    if use_fping; then
        hosts_family
        if [ $? -eq 99 ]; then
            ocf_log err "host_list can contain only host with same IP versions for fping"
            exit $OCF_ERR_CONFIGURED
        fi
    fi

    return $OCF_SUCCESS
}

ping_reload_agent() {
    return $OCF_SUCCESS
}

fping_check() {
    p_exe=fping
    hosts_family
    case $? in
        6) p_exe=fping6 ;;
        99) ocf_log err "Ambiguous IP versions in host_list: '$OCF_RESKEY_host_list'"; exit $OCF_ERR_CONFIGURED;;
    esac

    active=0

    timeout=$(expr $OCF_RESKEY_timeout \* 1000 / $OCF_RESKEY_attempts)

    cmd="$p_exe -r $OCF_RESKEY_attempts -t $timeout -B 1.0 $OCF_RESKEY_options $OCF_RESKEY_host_list"
    fping_output=$($cmd 2>&1); rc=$?
    active=$(echo "$fping_output" | grep "is alive" | wc -l)

    case $rc in
        0)
            if [ $OCF_RESKEY_debug -gt 1 ]; then
                ping_conditional_log info "$fping_output"
            fi
            ;;
        1)
            for h in $(echo "$fping_output" | grep "is unreachable" | awk '{print $1}'); do
                ping_conditional_log warn "$h is inactive: $fping_output"
            done
            ;;
        *)
            ocf_log err "Unexpected result for '$cmd' $rc: $(echo "$fping_output" | tr '\n' ';')"
            ;;
    esac

    return $active
}

ping_check() {
    active=0
    for host in $OCF_RESKEY_host_list; do
        p_exe=ping

        case $(uname) in
            Linux) p_args="-n -q -W $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts";;
            Darwin) p_args="-n -q -t $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts -o";;
            FreeBSD) p_args="-n -q -t $OCF_RESKEY_timeout -c $OCF_RESKEY_attempts -o";;
            *) ocf_log err "Unknown host type: $(uname)"; exit $OCF_ERR_INSTALLED;;
        esac

        case "$host" in
            *:*) p_exe=ping6
        esac

        ping_output=$($p_exe $p_args $OCF_RESKEY_options $host 2>&1); rc=$?

        case $rc in
            0)
                active=$(expr $active + 1)
                if [ $OCF_RESKEY_debug -gt 1 ]; then
                    ping_conditional_log info "$ping_output"
                fi
                ;;
            1) ping_conditional_log warn "$host is inactive: $ping_output";;
            *) ocf_log err "Unexpected result for '$p_exe $p_args $OCF_RESKEY_options $host' $rc: $ping_output";;
        esac
    done
    return $active
}

ping_update() {

    if use_fping; then
        fping_check
        active=$?
    else
        ping_check
        active=$?
    fi

    score=$(expr $active \* $OCF_RESKEY_multiplier)
    attrd_updater -n "$OCF_RESKEY_name" -B "$score" -d "$OCF_RESKEY_dampen"
    rc=$?
    case $rc in
        0) ping_conditional_log debug "Updated $OCF_RESKEY_name = $score" ;;
        *) ocf_log warn "Could not update $OCF_RESKEY_name = $score: rc=$rc";;
    esac
    if [ $rc -ne 0 ]; then
        return $rc
    fi

    if [ -n "$OCF_RESKEY_failure_score" ] && [ "$score" -lt "$OCF_RESKEY_failure_score" ]; then
        ocf_log warn "$OCF_RESKEY_name is less than failure_score($OCF_RESKEY_failure_score)"
        return 1
    fi
    return 0
}

use_fping() {
    ocf_is_true "$OCF_RESKEY_use_fping" && have_binary fping;
}

# return values:
#     4 IPv4
#     6 IPv6
#     0 indefinite (i.e. hostname)
host_family() {
    case $1 in
        *[0-9].*[0-9].*[0-9].*[0-9]) return 4 ;;
        *:*) return 6 ;;
        *) return 0 ;;
    esac
}

# return values same as host_family plus
#     99 ambiguous families
hosts_family() {
    # For fping allow only same IP versions or hostnames
    family=0
    for host in $OCF_RESKEY_host_list; do
        host_family "$host"
        f=$?
        if [ $family -ne 0 ] && [ $f -ne 0 ] && [ $f -ne $family ] ; then
            family=99
            break
        fi
        [ $f -ne 0 ] && family=$f
    done
    return $family
}

integer=$(echo ${OCF_RESKEY_timeout} | egrep -o '[0-9]*')
case "${OCF_RESKEY_timeout}" in
    *[0-9]ms|*[0-9]msec) OCF_RESKEY_timeout=$(expr $integer / 1000);;
    *[0-9]m|*[0-9]min) OCF_RESKEY_timeout=$(expr $integer \* 60);;
    *[0-9]h|*[0-9]hr)  OCF_RESKEY_timeout=$(expr $integer \* 60 \* 60);;
    *) OCF_RESKEY_timeout=$integer;;
esac

if [ -z "${OCF_RESKEY_timeout}" ]; then
    if [ -n "$OCF_RESKEY_host_list" ]; then
        host_count=$(echo $OCF_RESKEY_host_list | awk '{print NF}')
        OCF_RESKEY_timeout=$(expr $OCF_RESKEY_CRM_meta_timeout / $host_count / $OCF_RESKEY_attempts)
        OCF_RESKEY_timeout=$(expr $OCF_RESKEY_timeout / 1100) # Convert to seconds and finish 10% early
    else
        OCF_RESKEY_timeout=5
    fi
fi

if [ ${OCF_RESKEY_timeout} -lt 1 ]; then
    OCF_RESKEY_timeout=5
elif [ ${OCF_RESKEY_timeout} -gt 1000 ]; then
    # ping actually complains if this value is too high, 5 minutes is plenty
    OCF_RESKEY_timeout=300
fi

if [ "${OCF_RESKEY_CRM_meta_globally_unique}" = "false" ]; then
    : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/ping-${OCF_RESKEY_name}"}
else
    : ${OCF_RESKEY_pidfile:="${HA_VARRUN%%/}/ping-${OCF_RESOURCE_INSTANCE}"}
fi

# Check the debug option
case "${OCF_RESKEY_debug}" in
    true|True|TRUE|1)    OCF_RESKEY_debug=1;;
    false|False|FALSE|0) OCF_RESKEY_debug=0;;
    verbose|Verbose|VERBOSE|2) OCF_RESKEY_debug=2;;
    *)
        ocf_log warn "Value for 'debug' is incorrect. Please specify 'true', 'false', or 'verbose', not: ${OCF_RESKEY_debug}"
        OCF_RESKEY_debug=false
        ;;
esac

case "$__OCF_ACTION" in
meta-data)      meta_data
                exit $OCF_SUCCESS
                ;;
start)          ping_start;;
stop)           ping_stop;;
monitor)        ping_monitor;;
validate-all)   ping_validate;;
reload-agent)   ping_reload_agent;;
usage|help)     ping_usage
                exit $OCF_SUCCESS
                ;;
*)              ping_usage
                exit $OCF_ERR_UNIMPLEMENTED
                ;;
esac
exit $?

# vim: set filetype=sh expandtab tabstop=4 softtabstop=4 shiftwidth=4 textwidth=80: