注意: 本次安装操作系统环境为Anolis8.9(Centos7和Centos8应该也一样)
DataSophon版本为DDP-1.2.1
整合的安装包我放网盘了:
通过网盘分享的文件:impala-4.4.1.tar.gz等2个文件
链接: https://pan.baidu.com/s/18KfkO_BEFa5gVcc16I-Yew?pwd=za4k 提取码: za4k
1. Apache Impala的版本我选择的是目前Github上的最新版本4.4.1
Github上 Impala提供了rpm和deb两种安装方式,由于编译源码涉及到python包以及C++包机器难下载,所以我选择了Github提供的rpm包(注意RPM包中没有包含shell目录 deb包中是有的,如果使用rpm包的话,需要手动将这个shell文件夹拷贝进来).
2. 首先下载apache-impala-4.4.1-RELEASE_hive-3.1.3-x86_64.el8.8.rpm
将rpm包移动到/opt目录下,然年执行yum安装
wget https://github.com/apache/impala/releases/download/4.4.1/apache-impala-4.4.1-RELEASE_hive-3.1.3-x86_64.el8.8.rpm
mv apache-impala-4.4.1-RELEASE_hive-3.1.3-x86_64.el8.8.rpm /opt
yum install -y apache-impala-4.4.1-RELEASE_hive-3.1.3-x86_64.el8.8.rpm
记得安装好之后将shell文件夹拷贝进来啊
3. 安装之后得到/opt/impala文件夹,重命名为/opt/impala-4.4.1
cd /opt
mv impala impala-4.4.1
4. 修改conf/impala-env.sh脚本命令,修改以下三个的值
: ${JAVA_HOME:=/usr/local/jdk}
# Specify extra CLASSPATH.
: ${CLASSPATH:=${IMPALA_HOME}/conf/:${IMPALA_HOME}/lib/jars/*}
# Specify extra LD_LIBRARY_PATH.
: ${LD_LIBRARY_PATH:=${IMPALA_HOME}/lib/native/:${JAVA_HOME}/jre/lib/amd64/server/}
5. 在conf下将Hadoop的core-site.xml hdfs-site.xml以及hive的core-site.xml拷贝进来
hive-site.xml内容我只放了这些
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<configuration>
<property>
<name>hive.metastore.warehouse.dir</name>
<value>/user/hive/warehouse</value>
</property>
<property>
<name>hive.metastore.port</name>
<value>9083</value>
</property>
<property>
<name>hive.metastore.uris</name>
<value>thrift://node01:9083</value>
</property>
<property>
<name>hive.metastore.event.db.notification.api.auth</name>
<value>false</value>
</property>
<property>
<name>hive.metastore.dml.events</name>
<value>true</value>
</property>
<property>
<name>hive.metastore.transactional.event.listeners</name>
<value>org.apache.hive.hcatalog.listener.DbNotificationListener</value>
</property>
</configuration>
6. 打包tar.gz,并拷贝到DDP/packages目录下以及生成md5文件
cd /opt
# 打包tar.gz
tar -zcvf impala-4.4.1.tar.gz impala-4.4.1
# 将tar.gz包拷贝到DDP/packages目录下
cp impapa-4.4.1.tar.gz /opt/datasophon/DDP/packages
cd /opt/datasophon/DDP/packages
# 生成md5文件
java -jar file-md5-1.0-SNAPSHOT-jar-with-dependencies.jar impala-4.4.1.tar.gz
7. 编写IMPALA的service_ddl.json配置文件
cd /opt/datasophon/datasophon-manager-1.2.1/conf/meta/DDP-1.2.1/
mkdir IMPALA
cd IMPALA
# 创建json文件并写入内容
vi service_ddl.json
{
"name": "IMPALA",
"label": "Impala",
"description": "MPP(大规模并行处理)SQL查询引擎",
"version": "4.4.1",
"sortNum": 22,
"dependencies": ["HDFS", "HIVE"],
"packageName": "impala-4.4.1.tar.gz",
"decompressPackageName": "impala-4.4.1",
"roles": [{
"name": "StateStored",
"label": "StateStored",
"roleType": "master",
"runAs": {
"user": "impala",
"group": "hadoop"
},
"cardinality": "1+",
"sortNum": 1,
"logFile": "/var/log/impala/statestored.INFO",
"jmxPort": 2191,
"startRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["start", "statestored"]
},
"stopRunner": {
"timeout": "600",
"program": "bin/impala.sh",
"args": ["stop", "statestored"]
},
"statusRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["status", "statestored"]
},
"restartRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["restart", "statestored"]
},
"externalLink": {
"name": "StateStored Ui",
"label": "StateStored Ui",
"url": "http://${host}:25010"
}
}, {
"name": "Catalogd",
"label": "Catalogd",
"roleType": "master",
"runAs": {
"user": "impala",
"group": "hadoop"
},
"cardinality": "1+",
"sortNum": 2,
"logFile": "/var/log/impala/catalogd.INFO",
"jmxPort": 2191,
"startRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["start", "catalogd"]
},
"stopRunner": {
"timeout": "600",
"program": "bin/impala.sh",
"args": ["stop", "catalogd"]
},
"statusRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["status", "catalogd"]
},
"restartRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["restart", "catalogd"]
},
"externalLink": {
"name": "Catalogd Ui",
"label": "Catalogd Ui",
"url": "http://${host}:25020"
}
}, {
"name": "Impalad",
"label": "Impalad",
"roleType": "worker",
"runAs": {
"user": "impala",
"group": "hadoop"
},
"cardinality": "1+",
"sortNum": 3,
"logFile": "/var/log/impala/impalad.INFO",
"jmxPort": 2191,
"startRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["start", "impalad", "--enable_legacy_avx_support"]
},
"stopRunner": {
"timeout": "600",
"program": "bin/impala.sh",
"args": ["stop", "impalad"]
},
"statusRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["status", "impalad"]
},
"restartRunner": {
"timeout": "60",
"program": "bin/impala.sh",
"args": ["restart", "impalad", "--enable_legacy_avx_support"]
}
}],
"configWriter": {
"generators": [{
"filename": "statestored_flags",
"configFormat": "properties",
"outputDirectory": "conf",
"includeParams": ["-hostname", "-log_dir", "-minidump_path", "custom.statestored_flags"]
}, {
"filename": "catalogd_flags",
"configFormat": "properties",
"outputDirectory": "conf",
"includeParams": ["-hostname", "-state_store_host", "-log_dir", "-minidump_path", "custom.catalogd_flags"]
}, {
"filename": "impalad_flags",
"configFormat": "properties",
"outputDirectory": "conf",
"includeParams": ["-hostname", "-state_store_host", "-catalog_service_host", "-log_dir", "-minidump_path", "-mem_limit", "custom.impalad_flags"]
}]
},
"parameters": [{
"name": "-hostname",
"label": "impalad部署节点IP",
"description": "impalad部署节点IP",
"required": true,
"type": "input",
"value": "${host}",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "${host}"
}, {
"name": "-catalog_service_host",
"label": "catalog_service_host部署节点IP",
"description": "catalog_service_host部署节点IP",
"required": true,
"type": "input",
"value": "node01",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "${host}"
}, {
"name": "-state_store_host",
"label": "statestore部署节点IP",
"description": "statestore部署节点IP",
"required": true,
"type": "input",
"value": "node01",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "${host}"
}, {
"name": "-log_dir",
"label": "log_dir日志路径",
"description": "log_dir日志路径",
"required": true,
"type": "input",
"value": "/var/log/impala",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "/var/log/impala"
}, {
"name": "-minidump_path",
"label": "minidump_path路径",
"description": "minidump_path路径",
"required": true,
"type": "input",
"value": "/var/log/impala/minidumps",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "/var/log/impala/minidumps"
}, {
"name": "-mem_limit",
"label": "mem_limit",
"description": "mem_limit",
"required": true,
"type": "input",
"value": "80%",
"configurableInWizard": true,
"hidden": false,
"defaultValue": "80%"
}, {
"name": "custom.statestored_flags",
"label": "自定义配置statestored_flags",
"description": "自定义配置",
"configType": "custom",
"required": true,
"type": "multipleWithKey",
"value": [{
"-v": "1"
},
{
"-log_filename": "statestored"
},
{
"-max_log_files": "10"
},
{
"-max_log_size": "200"
}
],
"configurableInWizard": true,
"hidden": false,
"defaultValue": ""
}, {
"name": "custom.catalogd_flags",
"label": "自定义配置catalogd_flags",
"description": "自定义配置",
"configType": "custom",
"required": true,
"type": "multipleWithKey",
"value": [{
"-v": "1"
},
{
"-log_filename": "catalogd"
},
{
"-max_log_files": "10"
},
{
"-max_log_size": "200"
}
],
"configurableInWizard": true,
"hidden": false,
"defaultValue": ""
}, {
"name": "custom.impalad_flags",
"label": "自定义配置impalad_flags",
"description": "自定义配置",
"configType": "custom",
"required": true,
"type": "multipleWithKey",
"value": [{
"-v": "1"
},
{
"-log_filename": "impalad"
},
{
"-max_log_files": "10"
},
{
"-max_log_size": "200"
},
{
"-scratch_dirs": "/data/impala/impalad"
}
],
"configurableInWizard": true,
"hidden": false,
"defaultValue": ""
}]
}
8. 重启datasophon的api服务
cd /opt/datasophon/datasophon-manager-1.2.1/
启动:sh bin/datasophon-api.sh start api
停止:sh bin/datasophon-api.sh stop api
重启:sh bin/datasophon-api.sh restart api
9. 回到网页中就可以进行服务的安装了(注意安装过程中修改catalogd以及satastored所在的服务器的ip或者host名称)
由于我的目录设置为/var/log/impala, impala并没有权限进行创建该目录,只能手动创建(这个问题我知道咋解决),这个步骤要在安装服务之前就做吧.
mkdir -p /var/log/impala/minidumps
chmod 777 /var/log/impala
chmod 777 var/log/impala/minidumps
如果有读者想自己打包安装可以参考以下博客
编译部署apache-impala | 子崖说
Datasophon集成impala | 子崖说
https://zhuanlan.zhihu.com/p/348344999