@echo off
rem ***************************************
rem *            forcestop.bat            *
rem ***************************************

rem OCI CLI PATH
set OCIPath="C:\Users\opc\bin\oci.exe" ★ OCIコマンドパス

rem node1 setting info
set HOST1=server1 ★ サーバー１のホスト名
set INSTANCE_OCID_HOST1=ocid1.instance.oc1.ap-tokyo-1.xxxx ★ サーバー１のINSTANCE OCID
rem node2 setting info
set HOST2=server2 ★ サーバー２のホスト名
set INSTANCE_OCID_HOST2=ocid1.instance.oc1.ap-tokyo-1.xxxx ★ サーバー２のINSTANCE OCID

rem instance exec stop count
set STOP_LOOP_MAX=2

echo "START"
echo "DOWN SERVER NAME : %CLP_SERVER_DOWN%"
echo "LOCAL SERVER NAME: %CLP_SERVER_LOCAL%"

if %CLP_SERVER_DOWN%==%HOST1% (
    set INSTANCE_OCID=%INSTANCE_OCID_HOST1%
) else if %CLP_SERVER_DOWN%==%HOST2% (
    set INSTANCE_OCID=%INSTANCE_OCID_HOST2%
) else (
    echo "SERVER is not found."
    exit 1
)

rem stop instance
set=STOP_LOOP_COUNT=0

:STOP_LOOP

%OCIPath% compute instance action --action STOP --instance-id %INSTANCE_OCID%
set ret=%ERRORLEVEL%

if %ERRORLEVEL%==0 (
    echo "succeeded to stop instance. (%INSTANCE_OCID%)"
    GOTO STOP_CHECK
)

echo "failed to stop instance. (%INSTANCE_OCID%, ret=%ret%)"

timeout /t 1
set /a STOP_LOOP_COUNT=STOP_LOOP_COUNT+1

if %STOP_LOOP_COUNT%==%STOP_LOOP_MAX% (
    echo "stop instance retry over."
    echo "EXIT %ret%"
    exit %ret%
)

GOTO STOP_LOOP


:STOP_CHECK
rem server down check

:CHECK_LOOP

%OCIPath% compute instance get --instance-id %INSTANCE_OCID% | findstr "lifecycle-state" | findstr "STOPPED"
set ret=%ERRORLEVEL%

if %ERRORLEVEL%==0 (
    echo "%INSTANCE_OCID% has been stopped."
    GOTO EXIT
)
timeout /t 1

GOTO CHECK_LOOP

:EXIT
echo "EXIT 0"
exit 0

#! /bin/sh
#***********************************************
#*                forcestop.sh                 *
#***********************************************

OCIPath="/usr/bin/oci" ★OCIコマンドのパス

HOST1="server1" ★サーバー1のホスト名
INSTANCE_OCID_HOST1="ocid1.instance.oc1.ap-tokyo-1.xxxx" ★ サーバー１のINSTANCE OCID

HOST2="server2" ★サーバー2のホスト名
INSTANCE_OCID_HOST2="ocid1.instance.oc1.ap-tokyo-1.xxxx" ★ サーバー２のINSTANCE OCID

STOP_LOOP_MAX=2

if [ "${CLP_SERVER_DOWN}" = "${HOST1}" ]; then
    INSTANCE_OCID="${INSTANCE_OCID_HOST1}"
elif [ "${CLP_SERVER_DOWN}" = "${HOST2}" ]; then
    INSTANCE_OCID="${INSTANCE_OCID_HOST2}"
else
    echo "DOWN SERVER is not found."
    echo "EXIT 1"
    exit 1
fi

# インスタンスの停止要求
STOP_LOOP_COUNT=0

while [ "${STOP_LOOP_COUNT}" -lt "${STOP_LOOP_MAX}" ]
do
    sudo ${OCIPath} compute instance action --action STOP --instance-id ${INSTANCE_OCID}
    ret=$?
    if [ "${ret}" -eq 0 ]; then
        echo "succeeded to stop instance. (${CLP_SERVER_DOWN})"
        break
    fi
    echo "failed to stop instance. (${CLP_SERVER_DOWN}, ret=${ret})"
    sleep 1
    let STOP_LOOP_COUNT=${STOP_LOOP_COUNT}+1
done

if [ "${ret}" -ne 0 ]; then
    echo "stop instance retry over."
    echo "EXIT ${ret}"
    exit ${ret}
fi

# サーバーダウン確認
# サーバーの停止は強制停止スクリプトのタイムアウト時間確認される
# タイムアウト時間超過後は、強制停止スクリプトが失敗と判断される

while [ 0 ]
do
    STATUS=`sudo ${OCIPath} compute instance get --instance-id ${INSTANCE_OCID} | grep lifecycle-state | awk -F\" '{print $4}'`
    if [ $? -ne 0 ]; then
        echo "oci cli command error."
    else
        if [ "${STATUS}" = "STOPPED" ]; then
            echo "${CLP_SERVER_DOWN} has been stopped "
            break
        fi
        sleep 1
    fi
done

echo "EXIT 0"
exit 0

OSの起動時間 >= 強制停止スクリプトのタイムアウト + ハートビートタイムアウト + ハートビートインターバル

CLUSTERPROサービス起動時の遅延時間 >= 強制停止スクリプトのタイムアウト + ハートビートタイムアウト + ハートビートインターバル

Jul 22 02:52:35 server2 clusterpro: [I] Server server1 has been stopped. ★ 現用系サーバーのダウンを検出
Jul 22 02:52:36 server2 clusterpro: [I] Script for forced stop has started. ★ 強制停止スクリプトを実行開始
Jul 22 02:52:59 server2 clusterpro: [I] Script for forced stop has completed. ★ 強制停止スクリプトの実行完了
Jul 22 02:52:59 server2 clusterpro: [I] Failover group failover has started. ★ フェールオーバーグループの起動を開始
Jul 22 02:52:59 server2 clusterpro: [I] Activating group failover has started.
Jul 22 02:52:59 server2 clusterpro: [I] Activating exec resource has started.

[root@server2 ~]# oci compute instance get --instance-id ocid1.instance.oc1.ap-tokyo-1.xxxxx | grep lifecycle-state | awk -F\" '{print $4}'
STOPPED

Jul 22 07:51:24 server1 clusterpro: [E] Group failover has started on more than one server.
Jul 22 07:51:24 server1 clusterpro: [I] Since server server2 is not specified as that which suppresses shutdown at both-system activation detection, it executed the shutdown request.
Jul 22 07:51:24 server1 clusterpro: [I] Since server server1 is not specified as that which suppresses shutdown at both-system activation detection, it executed the shutdown request.