Alluxio 内存存储系统部署

一、文件下载和解压

1）下载地址：http://www.alluxio.org/download

2） 解压命令如下：

$ wget http://alluxio.org/downloads/files/1.2.0/alluxio-1.2.0-bin.tar.gz
$ tar xvfz alluxio-1.2.0-bin.tar.gz
$ cd alluxio-1.2.0

二、 配置文件更改

目前只是基本配置更改：

1） /data/spark/software/alluxio-1.2.0/conf下的  alluxio-env.sh.template 复制一份为： alluxio-env.sh

更改如下：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

#!/usr/bin/env bash # # The Alluxio Open Foundation licenses this work under the Apache License, version 2.0 # (the "License"). You may not use this work except in compliance with the License, which is # available at www.apache.org/licenses/LICENSE-2.0 # # This software is distributed on an "AS IS" basis, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, # either express or implied, as more fully set forth in the License. # # See the NOTICE file distributed with this work for information regarding copyright ownership. # # Copy it as alluxio-env.sh and edit that to configure Alluxio for your # site. This file is sourced to launch Alluxio servers or use Alluxio shell # commands. # # This file provides one way to configure Alluxio options by setting the # following listed environment variables. Note that, setting this file will not # affect jobs (e.g., Spark job or MapReduce job) that are using Alluxio client # as a library. Alternatively, you can edit alluxio-site.properties file, where # you can set all the configuration options supported by Alluxio # (http://alluxio.org/documentation/) which is respected by both external jobs # and Alluxio servers (or shell). # The directory where Alluxio deployment is installed. (Default: the parent directory of libexec/). export  ALLUXIO_HOME=/data/spark/software/alluxio-1.2.0 # The directory where log files are stored. (Default: ${ALLUXIO_HOME}/logs). # ALLUXIO_LOGS_DIR # Hostname of the master. # ALLUXIO_MASTER_HOSTNAME export ALLUXIO_MASTER_HOSTNAME=spark29 # This is now deprecated. Support will be removed in v2.0 # ALLUXIO_MASTER_ADDRESS #export ALLUXIO_MASTER_ADDRESS=spark29 # The directory where a worker stores in-memory data. (Default: /mnt/ramdisk). # E.g. On linux,  /mnt/ramdisk for ramdisk, /dev/shm for tmpFS; on MacOS, /Volumes/ramdisk for ramdisk # ALLUXIO_RAM_FOLDER export ALLUXIO_RAM_FOLDER=/data/spark/software/alluxio-1.2.0/ramdisk # Address of the under filesystem address. (Default: ${ALLUXIO_HOME}/underFSStorage) # E.g. "/my/local/path" to use local fs, "hdfs://localhost:9000/alluxio" to use a local hdfs # ALLUXIO_UNDERFS_ADDRESS export ALLUXIO_UNDERFS_ADDRESS=hdfs://spark29:9000 # How much memory to use per worker. (Default: 1GB) # E.g. "1000MB", "2GB" # ALLUXIO_WORKER_MEMORY_SIZE export ALLUXIO_WORKER_MEMORY_SIZE=12GB # Config properties set for Alluxio master, worker and shell. (Default: "") # E.g. "-Dalluxio.master.port=39999" # ALLUXIO_JAVA_OPTS # Config properties set for Alluxio master daemon. (Default: "") # E.g. "-Dalluxio.master.port=39999" # ALLUXIO_MASTER_JAVA_OPTS # Config properties set for Alluxio worker daemon. (Default: "") # E.g. "-Dalluxio.worker.port=49999" to set worker port, "-Xms2048M -Xmx2048M" to limit the heap size of worker. # ALLUXIO_WORKER_JAVA_OPTS # Config properties set for Alluxio shell. (Default: "") # E.g. "-Dalluxio.user.file.writetype.default=CACHE_THROUGH" # ALLUXIO_USER_JAVA_OPTS

 

2）worker 下面的添加worker节点的地址

spark24

spark30

spark31

spark32

spark33

三 、主机配置更改

1）在家目录下更改 .bash_profile 添加一下内容：

export TACHYON_HOME=/data/spark/software/alluxio-1.2.0

PATH=$PATH:$HOME/bin:$HADOOP/bin:$JAVA_HOME/bin:$TACHYON_HOME/bin

2） 生效配置

 source .bash_profile

四 、Spark 添加依赖Jar

1、在所有的spark主机的spark安装目录下的conf目录下

   更改spark-env.sh 后面添加：export SPARK_CLASSPATH="/data/spark/software/spark-1.5.2-bin-hadoop2.6/lib/alluxio-core-client-spark-1.2.0-jar-with-dependencies.jar:$SPARK_CLASSPATH"

五 、分发到各个Worker节点上去

1、alluxio 软件：scp -r ./alluxio-1.2.0 spark30:/data/spark/software/

六、格式化和启动

1、进入到alluxio的安装目录下面的bin目录，执行命令： alluxio format  进行内存格式化。

2、启动集群：./alluxio-start.sh all

七、可能遇到问题

1、启动worker报错，报错内容：Pseudo-terminal will not be allocated because stdin is not a terminal.

更改：alluxioinalluxio-workers.sh  的44行内容

原始内容为：

nohup ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no -t ${worker} ${LAUNCHER}  

改成如下：
nohup ssh -o ConnectTimeout=5 -o StrictHostKeyChecking=no -tt  $ {worker}  ${LAUNCHER} 

2、如果启动报sudo相关命令错误，是因为启动用户未在sudoers里面，需要将用户添加到此文件中，添加方法搜下root位置，再后面添加即可。

内容如下：

root ALL=(ALL) ALL
spark ALL=(ALL) ALL

同时把此文件中的：#Defaults    requiretty 注释掉。

3、如果还报错，可以在启动master之后，一个一个节点去启动worker。

八、 官网安装说明

官网安装说明：http://www.alluxio.org/docs/master/cn/Running-Alluxio-on-a-Cluster.html  有中文的，可以看看。