6 Zookeeper 실행

playdata data engineering

1 Zookeeper myid 편집

# master 서버
cat /usr/local/zookeeper/data/myid

# worker1 서버
vim /usr/local/zookeeper/data/myid
2

# worker2 서버
vim /usr/local/zookeeper/data/myid
3

---

2 Zookeeper 실행

# master 서버
/usr/local/zookeeper/bin/zkServer.sh start

# worker1 서버
/usr/local/zookeeper/bin/zkServer.sh start

# worker2 서버
/usr/local/zookeeper/bin/zkServer.sh start

---

3 Zookeeper 상태 확인

# master 서버
/usr/local/zookeeper/bin/zkServer.sh status

# worker1 서버
/usr/local/zookeeper/bin/zkServer.sh status

# worker2 서버
/usr/local/zookeeper/bin/zkServer.sh status

---

4 master -> Zookeeper 초기화

# zookeeper 초기화
hdfs zkfc -formatZK

---

5 master -> HDFS ZKFC 초기화 확인

# zkCli 실행
/usr/local/zookeeper/bin/zkCli.sh

# Hadoop 클러스터 확인
ls /hadoop-ha

# [my-hadoop-cluster] 확인 후 quit 명령으로 종료

# 종료
quit

• 가장 하단에 ERROR가 발생해도 놀라지마세요.
  

---

6 master, worker1, worker2 -> Journalnode 실행

# master 서버
hdfs --daemon start journalnode
jps # 결과 확인

# worker1 서버
hdfs --daemon start journalnode
jps # 결과 확인

# worker2 서버
hdfs --daemon start journalnode
jps # 결과 확인

• QuoromPeerMain : zookeeper가 동작하고 있다는 것을 의미
• JournalNode : journalnode가 동작하는 것을 의미