一、文件下載和解壓
1）下載地址：http://www.alluxio.org/download
2） 解壓命令如下：
$ 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
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: ${ALLUXIO_HOME}/underFSStorage)
# E.g. "/my/local/path" to use local fs, "hdfs://localhost:9000/alluxio" to use a local hdfs
# ALLUXIO_UNDERFS_ADDRESS
export
ALLUXIO_UNDERFS_ADDRESS=hdfs:
//spark29
:9000
# How much memory to use per worker. (Default: 1GB)
# E.g. "1000MB", "2GB"
# ALLUXIO_WORKER_MEMORY_SIZE
export
ALLUXIO_WORKER_MEMORY_SIZE=12GB
# Config properties set for Alluxio master, worker and shell. (Default: "")
# E.g. "-Dalluxio.master.port=39999"
# ALLUXIO_JAVA_OPTS
# Config properties set for Alluxio master daemon. (Default: "")
# E.g. "-Dalluxio.master.port=39999"
# ALLUXIO_MASTER_JAVA_OPTS
# Config properties set for Alluxio worker daemon. (Default: "")
# E.g. "-Dalluxio.worker.port=49999" to set worker port, "-Xms2048M -Xmx2048M" to limit the heap size of worker.
# ALLUXIO_WORKER_JAVA_OPTS
# Config properties set for Alluxio shell. (Default: "")
# E.g. "-Dalluxio.user.file.writetype.default=CACHE_THROUGH"
# ALLUXIO_USER_JAVA_OPTS
|
三 、主機配置更改
1）在家目錄下更改 .bash_profile 添加一下內容：
四 、Spark 添加依賴Jar
1、在所有的spark主機的spark安裝目錄下的conf目錄下
更改spark-env.sh 后面添加：export SPARK_CLASSPATH="/data/spark/software/spark-1.5.2-bin-hadoop2.6/lib/alluxio-core-client-spark-1.2.0-jar-with-dependencies.jar:$SPARK_CLASSPATH"
五 、分發到各個Worker節點上去
六、格式化和啟動
1、進入到alluxio的安裝目錄下面的bin目錄，執行命令： alluxio format 進行內存格式化。
2、啟動集群：./alluxio-start.sh all
七、可能遇到問題
1、啟動worker報錯，報錯內容：Pseudo-terminal will not be allocated because stdin is not a terminal.
更改：alluxio\bin\alluxio-workers.sh 的44行內容
原始內容為：
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、如果啟動報sudo相關命令錯誤，是因為啟動用戶未在sudoers里面，需要將用戶添加到此文件中，添加方法搜下root位置，再后面添加即可。
內容如下：
root ALL=(ALL) ALL
spark ALL=(ALL) ALL
同時把此文件中的：#Defaults requiretty 注釋掉。
3、如果還報錯，可以在啟動master之后，一個一個節點去啟動worker。
八、 官網安裝說明
官網安裝說明：http://www.alluxio.org/docs/master/cn/Running-Alluxio-on-a-Cluster.html 有中文的，可以看看。