본문 바로가기
엔지니어/Linux

How to make High Available Load Balancer(L4/L7) with haproxy and Pacemaker - 3/4

louie0 2016. 7. 6. 17:55
728x90
반응형

4. High Availability "Pacemaker" Install & Configuration


Active/Standby Node에 아래와 같은 과정을 통해 Pacemaker를 설치한다.


  • Install libraries
    • # yum update
    • # yum install libtool-ltdl.x86_64 openhpi-libs.x86_64 libibverbs.x86_64 librdmacm.x86_64



  • Install
    • # rpm -Uvh ./*.rpm




  • make /etc/ha.d/authkeys
    • # chmod 600 /etc/ha.d/authkeys
    • ClusterPasswordHere에는 Cluster(Active/Standby Node)가 인증에 사용할 적절한 Password를 입력한다.

    auth 2
1 crc
2 sha1 ClusterPasswordHere


  • make /etc/ha.d/ha.cf
    • Pacemaker는 Cluster Communication(Health Check/Configuration 등)을 위해 Messaging Layer로 heartbeat 또는 corosync를 사용한다. 본 문서에서는 heartbeat을 사용하였다.

    # Active Node
    logfile /var/log/ha-log

keepalive 1 # 1초 간격으로 Health Check.
deadtime 3  # 3초 안에 Health Check에 대한 응답이 없으면 상대 Node Failed로 판단.
ucast eth0 192.168.0.12 # 상대편 Node의 IP Address
crm yes     # Pacemaker 실행

node ACT
node STD

    # Standby Node
    logfile /var/log/ha-log

keepalive 1
deadtime 3
ucast eth0 192.168.0.11
crm yes

node ACT
node STD


  • edit /etc/hosts

    # Do not remove the following line, or various programs
# that require network functionality will fail.
127.0.0.1localhost.localdomain localhost

192.168.0.11 ACT
192.168.0.12 STD


  • Start Pacemaker
    • Active/Standby Node 양쪽 모두에서 Pacemaker를 실행한다.

    # /etc/init.d/heartbeat start
Starting High-Availability services:                       [  OK  ]



  • Pacemaker Configuration
    • Cluster Node 중 한곳에서만 설정하면 전체 Cluster에 적용된다.
 
        # crm
crm(live)# configure
crm(live)configure# property stonith-enabled=false
crm(live)configure# property no-quorum-policy=ignore
crm(live)configure# rsc_defaults resource-stickiness=100
crm(live)configure# rsc_defaults failure-timeout=0
  -> Resource가 다른 Node로 Failover 됐을 때 원래 Node로 Auto failback 하지 않도록(수동 failback) 설정
crm(live)configure# rsc_defaults migration-threshold=10
  -> Resource의 강제 migration을 결정하는 임계값 설정.
crm(live)configure# primitive VIP_192.168.0.10 ocf:heartbeat:IPaddr params ip=192.168.0.10 cidr_netmask=24 op monitor interval=5s
  -> Cluster는 Resource(VIP)의 상태를 5초 간격으로 Health Check 한다. Timeout 기본값은 15초.
  -> Health Check Failed이면 fail-count를 횟수만큼 증가 시킨다.
  -> fail-count > migration-threshold이면 Node의 상태에 상관없이 다른 Node에서 Resource를 Failover 한다.
crm(live)configure# primitive haproxy lsb:haproxy op monitor interval=5s
  -> haproxy도 Pacemaker의 Resource로 등록. Health Check Interval은 5초.
crm(live)configure# group Service_Name VIP haproxy
  -> VIP와 haproxy를 하나의 Resource로 Grouping.
crm(live)configure# location Service_Name_on_ACT Service_Name 100: ACT
  -> Resource의 Preference Node를 지정, 여기서는 Active Node. 
crm(live)configure# verify
  -> Configuration 검증.
crm(live)configure# commit
  -> Configuration 적용.

      

  • Cluster 상태 확인
 
      # crm_mon -1nfr
============
Last updated: Wed Feb 20 18:53:12 2013
Stack: Heartbeat
Current DC: ACT (07e78f2b-c4cc-42e6-b058-d7e0bd3a91e5)	- partition with quorum
Version: 1.0.12-unknown
2 Nodes configured, unknown expected votes
1 Resources configured.
============

Node ACT (07e78f2b-c4cc-42e6-b058-d7e0bd3a91e5): online
        haproxy (lsb:haproxy) Started
        VIP_192.168.0.10        (ocf::heartbeat:IPaddr) Started
Node STD (7b640fc0-0089-4db9-8c83-e34476c5e065): online

Inactive resources:


Migration summary:
* Node ACT:
* Node STD:
반응형

'엔지니어 > Linux' 카테고리의 다른 글

How to make High Available Load Balancer(L4/L7) with haproxy and Pacemaker - 1/4  (0) 2016.07.06
How to make High Available Load Balancer(L4/L7) with haproxy and Pacemaker - 2/4  (0) 2016.07.06
How to make High Available Load Balancer(L4/L7) with haproxy and Pacemaker - 4/4  (0) 2016.07.06
Howto install GlusterFS 3.5.1 on CentOS 6.5  (1) 2016.05.27
Iptable QoS(DSCP)  (0) 2016.05.27

'엔지니어/Linux' Related Articles

티스토리툴바