Document Type | Technical Information

CATEGORY | Installation

Applicable Product Version | Tibero 7FS04PS

Document Number | PINTI009

Overview

This document describes how to configure CM Failover in a unidirectional synchronization setup from Source DB (TAC) to Target DB (TAC).

This document covers "Part 3: CM Failover Setup".

Part 1: Prosync Installation Preparation and Agent Installation
Part 2: Instance Installation
Part 3: CM Failover Setup
Part 4: CM Failover Testing

Test Environment

Category	OS	IP	DBMS Version	Prosync Version
Source_TAC0	Rocky Linux release 8.10	10.10.10.61	Tibero 7.2.4 (build 305455)	4.6.0 (build 308985)
Source_TAC1	Rocky Linux release 8.10	10.10.10.62	Tibero 7.2.4 (build 305455)	4.6.0 (build 308985)
Target_TAC0	Rocky Linux release 8.10	10.10.10.63	Tibero 7.2.4 (build 305455)	4.6.0 (build 308985)
Target_TAC1	Rocky Linux release 8.10	10.10.10.64	Tibero 7.2.4 (build 305455)	4.6.0 (build 308985)

Note
The Prosync CM Failover feature is supported from Prosync version 4.3 and above.

Method

CM FAILOVER Setup

Stop DB & CM

Stop the tibero and cm processes on all servers.

$ tbdown immediate

Tibero instance terminated (IMMEDIATE mode).

$ tbcm -d
CM DOWN SUCCESS!

Modify $CM_SID.tip File

Add CM_ID at the very top line of $TB_HOME/config/$CM_SID.tip. The CM_ID is based on the AGENT_CM_ID configured in prs_install_agent.cfg during agent installation.

vi $TB_HOME/config/$CM_SID.tip

############ TAC (Tibero Active Cluster)
#### must

## node1
CM_ID=0 
CM_NAME=cm0
CM_UI_PORT=28629
CM_RESOURCE_FILE=/share/tibero_engine/config/cm0_res_file

### recommend
CM_HEARTBEAT_EXPIRE=300
CM_WATCHDOG_EXPIRE=290
CM_ENABLE_FAST_NET_ERROR_DETECTION=Y

Start CM & DB

Start cm and tibero processes on all servers.

$ tbcm -b
CM Guard daemon started up.
import resources from '/share/tibero_engine/config/cm0_res_file'...

TBCM 7.1.1 (Build 305455)

TmaxTibero Corporation Copyright (c) 2020-. All rights reserved.

Tibero cluster manager started up.
Local node name is (cm0:28629).


$ tbboot
Change core dump dir to /share/tibero_engine/bin/prof.
Listener port = 8629

Tibero 7

TmaxTibero Corporation Copyright (c) 2020-. All rights reserved.
Tibero instance started up (NORMAL mode).

Check CM

Use cmrctl show all to check the status.

$ cmrctl show all
Resource List of Node cm0
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net0       UP (private) 10.10.10.61/29000
     COMMON  cluster           cls0       UP inc: net0, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac0 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
=====================================================================
Resource List of Node cm1
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net1       UP (private) 10.10.10.62/29000
     COMMON  cluster           cls0       UP inc: net1, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac1 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
=====================================================================

Register CM GROUP

Add CM GROUP based on the information written in prs_install_agent.cfg during agent installation. Perform this on only one node.

<Syntax>
cmrctl add group --name <group_name> --cname <cluster_name> --grptype <type> --failover <true|false>

<Example>
$ cmrctl add group --name SRC_CM --cname cls0 --grptype prosync
Resource add success! (group, SRC_CM)

$ cmrctl add group --name TAR_CM --cname cls0 --grptype prosync
Resource add success! (group, TAR_CM)

Register CM AGENT

Add CM AGENT based on the information written in prs_install_agent.cfg during agent installation. Perform this on each node separately.

<Syntax>
cmrctl add agent --name <agent_name> --grpname <group_name> --script <directory_path> --pubnet <public_network_resource_name> --retry_cnt <retry_cnt>

<Example>
$ cmrctl add agent --name src_agent1 --grpname SRC_CM --script $PRS_HOME/bin/prs_0.sh 
Resource add success! (agent, src_agent1)

$ cmrctl add agent --name src_agent2 --grpname SRC_CM --script $PRS_HOME/bin/prs_1.sh 
Resource add success! (agent, src_agent2)

$ cmrctl add agent --name tar_agent1 --grpname TAR_CM --script $PRS_HOME/bin/prs_0.sh 
Resource add success! (agent, tar_agent1)

$ cmrctl add agent --name tar_agent2 --grpname TAR_CM --script $PRS_HOME/bin/prs_1.sh 
Resource add success! (agent, tar_agent2)

Verify CM Registration

Use cmrctl show all to check the registration status of GROUP & AGENT.

$ cmrctl show all
Resource List of Node cm0
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net0       UP (private) 10.10.10.61/29000
     COMMON  cluster           cls0       UP inc: net0, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac0 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         SRC_CM     DOWN type: prosync (failover: ON)
       cls0    agent     src_agent1     DOWN /share/prosync4/bin/prs_0.sh, start retry cnt: 0
=====================================================================
Resource List of Node cm1
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net1       UP (private) 10.10.10.62/29000
     COMMON  cluster           cls0       UP inc: net1, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac1 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         SRC_CM     DOWN type: prosync (failover: ON)
       cls0    agent     src_agent2     DOWN /share/prosync4/bin/prs_1.sh, start retry cnt: 0
=====================================================================

Note
If the agent is in DEACT status, check the $PRS_HOME/var/cmagent.log to verify the rc result value.
If rc is 127, modify the execution permission of $PRS_HOME/bin/prs_0.sh or change /bin/sh to /bin/bash to resolve the issue.
The command to reactivate the deact status is as follows:
$ cmrctl act agent --name <AGENT_ID>

Start CM GROUP

Start prosync using cm and start the cm group to use the failover feature.

$ cmrctl start group --name SRC_CM
=================================== SUCCESS! ===================================
 Succeeded to request at each node to boot resources under the group(SRC_CM).
 Please use "cmrctl show group --name SRC_CM" to verify the result.
================================================================================

$ cmrctl show all
Resource List of Node cm0
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net0       UP (private) 10.10.10.61/29000
     COMMON  cluster           cls0       UP inc: net0, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac0 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         SRC_CM       UP type: prosync (failover: ON)
       cls0    agent     src_agent1       UP /share/prosync4/bin/prs_0.sh, start retry cnt: 0
=====================================================================
Resource List of Node cm1
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net1       UP (private) 10.10.10.62/29000
     COMMON  cluster           cls0       UP inc: net1, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     source_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    source_tac1 UP(NRML) source_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         SRC_CM       UP type: prosync (failover: ON)
       cls0    agent     src_agent2       UP /share/prosync4/bin/prs_1.sh, start retry cnt: 0
=====================================================================

$ cmrctl start group --name TAR_CM
=================================== SUCCESS! ===================================
 Succeeded to request at each node to boot resources under the group(TAR_CM).
 Please use "cmrctl show group --name TAR_CM" to verify the result.
================================================================================

$ cmrctl show all
Resource List of Node cm0
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net0       UP (private) 10.10.10.63/29000
     COMMON  cluster           cls0       UP inc: net0, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     target_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    target_tac0 UP(NRML) target_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         TAR_CM       UP type: prosync (failover: ON)
       cls0    agent     tar_agent1       UP /share/prosync4/bin/prs_0.sh, start retry cnt: 0
=====================================================================
Resource List of Node cm1
=====================================================================
  CLUSTER     TYPE        NAME       STATUS           DETAIL
----------- -------- -------------- -------- ------------------------
     COMMON  network           net1       UP (private) 10.10.10.64/29000
     COMMON  cluster           cls0       UP inc: net1, pub: N/A
       cls0     file         cls0:0       UP /share/tac/CMFILE/CMFILE
       cls0  service     target_tac       UP Database, Active Cluster (auto-restart: OFF)
       cls0       db    target_tac1 UP(NRML) target_tac, /share/tibero_engine, failed retry cnt: 0
       cls0    group         TAR_CM       UP type: prosync (failover: ON)
       cls0    agent     tar_agent2       UP /share/prosync4/bin/prs_1.sh, start retry cnt: 0
=====================================================================

Verify prosync Startup

Check the prosync startup status using the prosync admin process.

$ prs_adm

ProSync 4 - Admin Utility

TmaxData Corporation Copyright (c) 2024-. All rights reserved.

Admin> status

prs_agent ID: src_agent1, HOST: 10.10.10.61, PORT: 7600, CM_GROUP: SRC_CM, CM_ID: 0 is running
prs_agent ID: src_agent2, HOST: 10.10.10.62, PORT: 7700, CM_GROUP: SRC_CM, CM_ID: 1 is running
prs_agent ID: tar_agent1, HOST: 10.10.10.63, PORT: 7800, CM_GROUP: TAR_CM, CM_ID: 0 is running
prs_agent ID: tar_agent2, HOST: 10.10.10.64, PORT: 7900, CM_GROUP: TAR_CM, CM_ID: 1 is running


Instance ID: [PRS_FAILOVER]
PRS_FAILOVER_ext1 (1) is running (prs_agent ID : src_agent1, HOST: 10.10.10.61, PORT: 7600)
PRS_FAILOVER_ext2 (2) is running (prs_agent ID : src_agent2, HOST: 10.10.10.62, PORT: 7700)
PRS_FAILOVER_apply1 (1) is running (prs_agent ID : tar_agent1, HOST: 10.10.10.63, PORT: 7800)
PRS_FAILOVER_llob (1) is running (prs_agent ID : src_agent1, HOST: 10.10.10.61, PORT: 7600)

Search

Welcome to Tibero GTS!

Prosync TAC to TAC Unidirectional Synchronization CM Failover Setup-3

Overview

Test Environment

Method

CM FAILOVER Setup

Stop DB & CM

Modify $CM_SID.tip File

Start CM & DB

Check CM

Register CM GROUP

Register CM AGENT

Verify CM Registration

Start CM GROUP

Verify prosync Startup

업무 외 시간 안내

Search

Welcome to Tibero GTS!

Overview

Test Environment

Method

CM FAILOVER Setup

Stop DB & CM

Modify $CM_SID.tip File

Start CM & DB

Check CM

Register CM GROUP

Register CM AGENT

Verify CM Registration

Start CM GROUP

Verify prosync Startup