Chinaunix首页 | 论坛 | 博客
  • 博客访问: 689372
  • 博文数量: 95
  • 博客积分: 1754
  • 博客等级: 上尉
  • 技术积分: 1607
  • 用 户 组: 普通用户
  • 注册时间: 2010-03-12 10:06
文章分类

全部博文(95)

文章存档

2015年(3)

2013年(15)

2012年(77)

分类:

2012-04-01 09:13:11

原文地址:HBASE安装和简单测试 作者:kerlion


1. 修改HDFSS设置
vi conf/hdfs-site.xml
增加下面的设置,HBASE需要访问大量的文件

dfs.datanode.max.xcievers
4096



2. 设置NTP同步
rpm -qa |grep ntp

master用缺省配置
slaves:
vi /etc/ntp.conf
server 192.168.110.127
把缺省配置的server都去掉,改为master的地址

chkconfig ntpd on
service ntpd restart


另外:最好使用同一时区
ln -sf /usr/share/zoneinfo/posix/Asia/Shanghai /etc/localtime

3. 修改nofile和nproc设置
HBase需要使用很多文件,每次flush都是写一个新文件,缺省1024远远不够

vi /etc/security/limits.conf
hadoop - nofile 32768
hadoop - nproc 32768

重新登录hadoop,验证一下
ulimit -a



4.下载和安装
到去下载最新的稳定版本
tar zxf hbase-0.92.1.tar.gz



5. 设置环境变量
export HBASE_HOME=$HOME/hbase-0.92.1
export HBASE_CONF_DIR=$HOME/hbase-conf

同时设置添加到PATH和CLASSPATH


6. 配置
cp -r $HBASE_HOME/conf $HOME/hbase-conf
vi hbase-env.sh

export JAVA_HOME=/usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/jre
export HBASE_HEAPSIZE=300
export HBASE_OPTS="-XX:+UseConcMarkSweepGC"
export HBASE_LOG_DIR=${HBASE_HOME}/logs
export HBASE_MANAGES_ZK=true

vi hdfs-site.xml
添加下面的配置,开启durable sync特性,Hadoop 0.20.205以上的版本有的功能
开启这个功能十分重要,否则HBASE会丢失数据。(本人猜测,这是些Hlog的需要,需要随时appdend redo log,而HDFS一般只能建新文件)


dfs.support.append
true



在Hadoop -.20.205的elease notes上有这么一句:
* This release includes a merge of append/hsynch/hflush capabilities from 0.20-append branch, to support HBase in secure mode.

7. 设置Fully-distributed模式
vi hdfs-site.xml
设置hbase.rootdir和hbase.cluster.distributed=true

hbase.rootdir
hdfs://master:9000/hbase


hbase.cluster.distributed
true



8. 设置RegionServers
cat regionservers
slave1
slave2


9. 配置ZooKeepers
vi hbase-env.sh
export HBASE_MANAGES_ZK=true

vi hdfs-site.xml

hbase.zookeeper.property.clientPort
2222


hbase.zookeeper.quorum
slave1,slave2


hbase.zookeeper.property.dataDir
/home/hadoop/zookeeper


10. 复制安装配置到其他节点

scp -r conf slave1:
scp -r conf slave2:

scp -r hbase-conf slave1:
scp -r hbase-conf slave2:
scp -r hbase-0.92.1 slave1:
scp -r hbase-0.92.1 slave2:
scp -r .bash_profile slave1:
scp -r .bash_profile slave2:



11. 重新登录,重启hadoop
stop-all.sh
start-all.sh
jps

12. 启动HBASE
start-hbase.sh

验证,

用jps命令查看java进程
Master上有
11420 HMaster
ZoomKeeper上有
575 HQuorumPeer
RegionServer上有
686 HRegionServer


13 简单测试
hbase shell
hbase(main):006:0> create 'test','data'
0 row(s) in 1.1190 seconds

hbase(main):007:0> list
TABLE
test
1 row(s) in 0.0270 seconds

hbase(main):009:0> put 'test','1','data:1','xxxx'
0 row(s) in 0.1220 seconds

hbase(main):010:0> put 'test','1','data:1','xxxx'
0 row(s) in 0.0120 seconds

hbase(main):011:0> put 'test','1','data:1','xxxx'
0 row(s) in 0.0120 seconds

hbase(main):015:0* put 'test','2','data:2','yyy'
0 row(s) in 0.0080 seconds

hbase(main):016:0> put 'test','3','data:3','zzz'
0 row(s) in 0.0070 seconds

hbase(main):017:0>
hbase(main):018:0*
hbase(main):019:0* scan 'test'
ROW COLUMN+CELL
1 column=data:1, timestamp=1333160616029, value=xxxx
2 column=data:2, timestamp=1333160650780, value=yyy
3 column=data:3, timestamp=1333160664490, value=zzz
3 row(s) in 0.0260 seconds

hbase(main):020:0>



14. 查看了HDFS上建立的文件
./hadoop dfs -lsr /hbase
Warning: $HADOOP_HOME is deprecated.

drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/-ROOT-
-rw-r--r-- 2 hadoop supergroup 551 2012-03-31 10:07 /hbase/-ROOT-/.tableinfo.0000000001
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/-ROOT-/.tmp
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/-ROOT-/70236052
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/-ROOT-/70236052/.oldlogs
-rw-r--r-- 2 hadoop supergroup 411 2012-03-31 10:07 /hbase/-ROOT-/70236052/.oldlogs/hlog.1333159627476
-rw-r--r-- 2 hadoop supergroup 109 2012-03-31 10:07 /hbase/-ROOT-/70236052/.regioninfo
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/-ROOT-/70236052/info
-rw-r--r-- 2 hadoop supergroup 714 2012-03-31 10:07 /hbase/-ROOT-/70236052/info/bd225e173164476f88111f622f5a7839
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.META.
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.META./1028785192
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.META./1028785192/.oldlogs
-rw-r--r-- 2 hadoop supergroup 124 2012-03-31 10:07 /hbase/.META./1028785192/.oldlogs/hlog.1333159627741
-rw-r--r-- 2 hadoop supergroup 111 2012-03-31 10:07 /hbase/.META./1028785192/.regioninfo
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.META./1028785192/info
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.logs
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.logs/slave1,60020,1333159627316
-rw-r--r-- 3 hadoop supergroup 293 2012-03-31 10:07 /hbase/.logs/slave1,60020,1333159627316/slave1%2C60020%2C1333159627316.1333159637444
-rw-r--r-- 3 hadoop supergroup 0 2012-03-31 10:07 /hbase/.logs/slave1,60020,1333159627316/slave1%2C60020%2C1333159627316.1333159637904
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:07 /hbase/.logs/slave2,60020,1333159627438
-rw-r--r-- 3 hadoop supergroup 0 2012-03-31 10:07 /hbase/.logs/slave2,60020,1333159627438/slave2%2C60020%2C1333159627438.1333159638583
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:18 /hbase/.oldlogs
-rw-r--r-- 2 hadoop supergroup 38 2012-03-31 10:07 /hbase/hbase.id
-rw-r--r-- 2 hadoop supergroup 3 2012-03-31 10:07 /hbase/hbase.version
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:22 /hbase/test
-rw-r--r-- 2 hadoop supergroup 513 2012-03-31 10:22 /hbase/test/.tableinfo.0000000001
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:22 /hbase/test/.tmp
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:22 /hbase/test/929f7e1caca5825974e0e991543fe2c5
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:22 /hbase/test/929f7e1caca5825974e0e991543fe2c5/.oldlogs
-rw-r--r-- 2 hadoop supergroup 124 2012-03-31 10:22 /hbase/test/929f7e1caca5825974e0e991543fe2c5/.oldlogs/hlog.1333160541983
-rw-r--r-- 2 hadoop supergroup 219 2012-03-31 10:22 /hbase/test/929f7e1caca5825974e0e991543fe2c5/.regioninfo
drwxr-xr-x - hadoop supergroup 0 2012-03-31 10:22 /hbase/test/929f7e1caca5825974e0e991543fe2c5/data


Error
==========================================
slave1: java.io.IOException: Could not find my address: db1 in list of ZooKeeper quorum servers
slave1: at org.apache.hadoop.hbase.zookeeper.HQuorumPeer.writeMyID(HQuorumPeer.java:133)
slave1: at org.apache.hadoop.hbase.zookeeper.HQuorumPeer.main(HQuorumPeer.java:60)
Reason:hostname是db1,但是我配置的名字是slave1,但是是同一个IP.HBase会用hostname取得的主机名来方向方向解析DNS
Solution:
修改hostname为slave1重新启动server




阅读(873) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~