OpenTSDB使用demo及其httpAPI的保护

淩亂°似流年 2022-02-05 11:21 370阅读 0赞

### 【目录】 ###

### 【第一章】Opentsdb运行环境准备 ###

### 【第二章】安装HA-hadoop集群、HA-HBASE集群、zookeeper集群 ###

### 【第三章】安装openTSDB ###

### 【第四章】通过Nginx对OpenTSDB做用户验证，确保opentsdb http接口安全 ###

###  ###

本文提到的各种服务和操作都经过本人验证，期间的坑已经替你们踩过了，读者可以参照本文档来一步步实施。觉得有用可以收藏。

### 【第一章】Opentsdb运行环境准备 ###

【操作系统】CentOS Linux release 7.2.1511 (Core)

【主机列表】

主机名 内网IP地址

opentsdb-1-4 192.168.35.2

opentsdb-1-5 192.168.35.3

opentsdb-1-1 192.168.35.4

opentsdb-1-3 192.168.35.5

opentsdb-1-2 192.168.35.6

【系统脚本语言版本】内置的Python 2.7.5

【opentsdb要求java版本Java-1.6+】

\[root@opentsdb-1-4 ~\]\# yum makecache \# 更新yum源

\[root@opentsdb-1-4 tony\_soft\]\# for x in \{5,1,3,2\}; do ssh root@opentsdb-1-$x yum makecache; done

\[root@opentsdb-1-4 tony\_soft\]\# yum install -y java-1.8.0-openjdk \# 使用yum安装JDK-1.8

解决no jps in (/usr/local/sbin问题：yum install -y java-1.8.0-openjdk-devel.x86\_64

继续为其他四个节点分别安装openjdk和jps：

\[root@opentsdb-1-4 tony\_soft\]\# for x in \{5,1,3,2\}; do ssh root@opentsdb-1-$x yum install -y java-1.8.0-openjdk java-1.8.0-openjdk-devel.x86\_64 >> /tmp/jdk\_and\_jps\_INSTALLING.log 2>&1;java -version && jps;echo "FINISHED\_\_\_\_ $x";done

【分别下载zookeeper-3.4.14、Hadoop-2.8.4、hbase-1.4.9】

wget [https://mirrors.tuna.tsinghua.edu.cn/apache/zookeeper/zookeeper-3.4.14/zookeeper-3.4.14.tar.gz][https_mirrors.tuna.tsinghua.edu.cn_apache_zookeeper_zookeeper-3.4.14_zookeeper-3.4.14.tar.gz]

wget [https://mirrors.tuna.tsinghua.edu.cn/apache/hadoop/common/hadoop-2.8.4/hadoop-2.8.4.tar.gz][https_mirrors.tuna.tsinghua.edu.cn_apache_hadoop_common_hadoop-2.8.4_hadoop-2.8.4.tar.gz]

wget [https://mirrors.tuna.tsinghua.edu.cn/apache/hbase/stable/hbase-1.4.9-bin.tar.gz][https_mirrors.tuna.tsinghua.edu.cn_apache_hbase_stable_hbase-1.4.9-bin.tar.gz]

【配置ssh免密登录】

确保从opentsdb-1-4到其他节点能ssh免密登录，具体实施可参考博文：

[https://jingyan.baidu.com/article/bea41d43b558ffb4c51be6db.html][https_jingyan.baidu.com_article_bea41d43b558ffb4c51be6db.html]

【开发自定义批处理脚本】

脚本：

可执行文件callall 用于同时在集群所有节点执行一条命令

可执行文件send2others 用于向其他节点(除opentsdb-1-4之外的节点)传送文件

都放到/usr/local/bin里面，给一个合适的权限。

\[root@opentsdb-1-4 etc\]\# ll -ht /usr/local/bin/ | head -n 3

total 8

\-rwxr-xr-x 1 root root 278 4月 24 15:57 callall

\-rwxr-xr-x 1 root root 384 4月 24 16:00 send2others

\[root@opentsdb-1-4 etc\]\# cat /usr/local/bin/callall

\#!/bin/bash

if \[\[ $\# -lt 1 \]\] ; then echo no params ; exit ; fi

cmd=$\*

for h in \{"opentsdb-1-4","opentsdb-1-5","opentsdb-1-1","opentsdb-1-3","opentsdb-1-2"\}

echo "正在$h 执行命令$cmd"

ssh $h "$cmd"

if \[ $? -eq 0 \];then

echo "--------命令执行成功！"

done ;

\[root@opentsdb-1-4 etc\]\# cat /usr/local/bin/send2others

\#!/bin/bash

if \[\[ $\# -lt 1 \]\] ; then echo no params ; exit ; fi

p=$1

dir=\`dirname $p\`

filename=\`basename $p\`

cd $dir

fullpath=\`pwd -P .\`

user=\`whoami\`

for h in \{"opentsdb-1-5","opentsdb-1-1","opentsdb-1-3","opentsdb-1-2"\}

echo "正在同步文件到$h======="

rsync -lr $p $\{user\}@$h:$fullpath

if \[ $? -eq 0 \];then

echo "--------已成功同步到$h"

done ;

【各节点安装NTP时间同步服务】

安装、配置可参考《CentOS 7 中使用NTP进行时间同步》[https://www.linuxidc.com/Linux/2015-11/124911.htm][https_www.linuxidc.com_Linux_2015-11_124911.htm]

【同步环境变量到其他四个节点】

创建符号链接：

callall ln -s hadoop-2.8.4 hadoop

callall ln -s hbase-1.4.9 hbase

callall ln -s zookeeper-3.4.14 zookeeper

send2others hbase && send2others zookeeper \# 同步软连接到其他节点的相应位置

callall 'ls -al /tony\_soft/'

先在/etc/profile配好JAVA\_HOME、HADOOP\_HOME、HBASE\_HOME、ZOOKEEPER\_HOME；再用send2others 同步到其他四个节点：

cd /etc && send2others profile && callall 'source /etc/profile'

callall echo $PATH && callall echo $HADOOP\_HOME && callall echo $HBASE\_HOME && callall echo $ZOOKEEPER\_HOME

send2others hadoop && send2others hbase && send2others zookeeper \# 同步符号链接到其他节点

send2others hbase-1.4.9 && send2others zookeeper-3.4.14 && send2others hadoop-2.8.4 \# 同步各个软件包到其他节点

### 【第二章】安装HA-hadoop集群、HA-HBASE集群、zookeeper集群 ###

【配置HA-hadoop集群】

角色规划：

opentsdb-1-4 NameNode, QPM(zk), Hmaster, ResourceManager

opentsdb-1-5 NameNode, QPM(zk), Hmaster, ResourceManager

opentsdb-1-1 JournalNode, QPM(zk), DataNode, HRegionServer

opentsdb-1-3 JournalNode, DataNode, HRegionServer

opentsdb-1-2 JournalNode, DataNode, HRegionServer

配置静态IP(略)、关闭防火墙：

callall service iptables stop，再callall cat /etc/selinux/config | grep SELINUX \# 确保每个主机的防火墙已关闭SELINUX=disabled

禁用Transparent Hugepage(为了提升hadoop集群性能)：

配置省略, 可参考[https://www.staroon.dev/2017/11/05/SetEnv/][https_www.staroon.dev_2017_11_05_SetEnv]）

为每个节点添加用户hadoop：

callall groupadd hadoop && callall useradd -g hadoop hadoop

创建hdfs相关目录(name目录、jn日志目录)并赋权：

callall 'mkdir -p **/dfs/data**' && callall 'chown -R hadoop:hadoop **/dfs**'

for i in \{1,3,2\}; do ssh opentsdb-1-$i mkdir **/jn\_edits**; done

for i in \{1,3,2\}; do ssh opentsdb-1-$i chown -R hadoop:hadoop **/jn\_edits**; done

拷贝hadoop相关命令到/usr/local/bin/目录下并赋权：

callall 'cp /tony\_soft/hadoop-2.8.4/bin/hdfs /usr/local/bin/ && chown hadoop:hadoop /usr/local/bin/hdfs'

callall ls -lh /usr/local/bin/hdfs

修改主要配置文件：

**core-site.xml**

<name>fs.defaultFS</name>

</property>

<name>fs.trash.interval</name>

</property>

<name>ha.zookeeper.quorum</name>

<value>opentsdb-1-4:2181,opentsdb-1-5:2181,opentsdb-1-1:2181</value>

</property>

<name>dfs.namenode.name.dir</name>

<description>namenode存放数据的目录,所有权是hadoop:hadoop</description>

</property>

</configuration>

**hdfs-site.xml**

<name>dfs.nameservices</name>

</property>

<!—配置故障自动转移的实现方式 -->

<name>dfs.client.failover.proxy.provider.tony</name>

<value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>

</property>

<name>dfs.ha.namenodes.tony</name>

</property>

<name>dfs.namenode.rpc-address.tony.nn1</name>

<value>opentsdb-1-4:8020</value>

</property>

<name>dfs.namenode.http-address.tony.nn1</name>

<value>opentsdb-1-4:50070</value>

</property>

<name>dfs.namenode.rpc-address.tony.nn2</name>

<value>opentsdb-1-5:8020</value>

</property>

<name>dfs.namenode.http-address.tony.nn2</name>

<value>opentsdb-1-5:50070</value>

</property>

<name>dfs.namenode.shared.edits.dir</name>

<value>qjournal://opentsdb-1-1:8485;opentsdb-1-3:8485;opentsdb-1-2:8485/tony</value>

</property>

<name>dfs.journalnode.edits.dir</name>

<value>**/jn\_edits**</value>

</property>

<name>dfs.ha.automatic-failover.enabled</name>

</property>

<name>dfs.ha.fencing.methods</name>

<value>

sshfence

shell(/bin/true)

</value>

</property>

<name>dfs.ha.fencing.ssh.private-key-files</name>

</property>

<name>dfs.ha.fencing.ssh.connect-timeout</name>

</property>

<name>dfs.replication</name>

</property>

</configuration>

**mapred-site.xml**

<name>mapreduce.framework.name</name>

</property>

</configuration>

**yarn-site.xml**

<name>yarn.resourcemanager.ha.enabled</name>

</property>

<name>yarn.resourcemanager.cluster-id</name>

<value>tonyyarn</value>

</property>

<name>yarn.resourcemanager.ha.rm-ids</name>

</property>

<name>yarn.resourcemanager.hostname.rm1</name>

<value>opentsdb-1-4</value>

</property>

<name>yarn.resourcemanager.hostname.rm2</name>

<value>opentsdb-1-5</value>

</property>

<name>yarn.resourcemanager.webapp.address.rm1</name>

<value>opentsdb-1-4:8088</value>

</property>

<name>yarn.resourcemanager.webapp.address.rm2</name>

<value>opentsdb-1-5:8088</value>

</property>

<name>yarn.resourcemanager.zk-address</name>

<value>opentsdb-1-4:2181,opentsdb-1-5:2181,opentsdb-1-1:2181</value>

</property>

<name>yarn.nodemanager.aux-services</name>

<value>mapreduce\_shuffle</value>

</property>

</configuration>

**slaves**

opentsdb-1-1

opentsdb-1-3

opentsdb-1-2

【启动HDFS之前的准备】

send2others hadoop-env.sh \# hadoop-env.sh添加完JAVA\_HOME之后，同步到其他节点

cp /tony\_soft/hadoop-2.8.4/bin/hdfs /usr/local/bin/hdfs

cp /tony\_soft/hadoop-2.8.4/bin/hadoop /usr/local/bin/

cd /tony\_soft/hadoop-2.8.4/bin && send2others hdfs && send2others hadoop

\[root@opentsdb-1-4 hadoop\]\# ll /usr/local/bin/

\-rwxr-xr-x 1 root root 278 4月 24 15:57 callall

\-rwxr-xr-x 1 hadoop hadoop 6567 4月 24 19:44 hadoop

\-rwxr-xr-x 1 hadoop hadoop 12383 4月 24 19:43 hdfs

\-rwxr-xr-x 1 root root 389 4月 24 16:27 send2others

\-rwxr-xr-x 1 root root 6773 4月 25 13:52 zkServer.sh

以hadoop用户启动HDFS的前提是hadoop用户能ssh免密登录到其他节点

先生成密钥对：

su hadoop

ssh-keygen -t rsa \# 一路回车即可

以root用户在/home/hadoop/.ssh目录下：

callall 'mkdir -m 700 /home/hadoop/.ssh' \# 批量创建权限码为700的hadoop家目录

send2others id\_rsa.pub \# 同步公钥到其他hadoop家目录下

callall 'chmod 644 /home/hadoop/.ssh/authorized\_keys'

callall 'chown hadoop:hadoop /home/hadoop/.ssh/authorized\_keys'

callall 'chmod 600 /home/hadoop/.ssh/id\_rsa.pub'

分别切换到hadoop用户,将opentsdb-1-4机器上的hadoop用户下用ssh-keygen生成的

/home/hadoop/.ssh/id\_rsa.pub公钥分发到opentsdb-1-\{5,1,3,2\}机器上的/home/hadoop/.ssh/authorized\_keys

callall 'cat /home/hadoop/.ssh/id\_rsa.pub > /home/hadoop/.ssh/authorized\_keys'

分发后，验证是否可免密ssh登录：for h in \{5,1,3,2\}; do ssh opentsdb-1-$h hostname; done

【配置zookeeper集群】

ZK三个节点：opentsdb-1-4、opentsdb-1-5、opentsdb-1-1

cd $ZOOKEEPER\_HOME

callall mkdir -m 755 /zkdata;done

callall chown -R hadoop:hadoop /zkdata;done

callall ls -l / | grep --color zkdata;done

cp zoo\_sample.cfg zoo.cfg

vim zoo.cfg \#指定部署zookeeper的3个节点; 指定zk数据存放目录和日志存放目录

![2019050517532057.png][]

for i in \{3,4\}; do scp conf/zoo.cfg 192.168.35.$i:/tony\_soft/zookeeper-3.4.14/conf/ ; done \# 同步到其他两个节点

mkdir -m 755 zklog \# 在$ZOOKEEPER\_HOME目录下创建zk日志存放目录

chown -R hadoop:hadoop zklog

\# 为其他两个zk节点创建zk日志存放目录并改权限

for i in \{3,4\};do ssh 192.168.35.$i cd $ZOOKEEPER\_HOME && mkdir -m 755 zklog && chown -R hadoop:hadoop zklog; done

按照zoo.cfg配置的3个zk节点，分别在每个节点的/zkdata目录下创建myid文件：

opentsdb-1-4机器上执行cd /zkdata && echo 1 >> myid

opentsdb-1-5机器上执行cd /zkdata && echo 2 >> myid

opentsdb-1-1机器上执行cd /zkdata && echo 3 >> myid

for h in \{5,1\}; do ssh opentsdb-1-$h 'cp /tony\_soft/zookeeper-3.4.14/bin/zkServer.sh /usr/local/bin/'; done

for h in \{5,1\}; do ssh opentsdb-1-$h 'chown hadoop:hadoop /usr/local/bin/zkServer.sh'; done

【以hadoop用户按顺序启动zk集群、启动JN集群、格式化HDFS、格式化zkfc、启动hdfs集群】

\#必须用原生的**zkServer.sh启动**，如果用放在/usr/local/bin/zkServer.sh启动就会报各种错

\[hadoop@opentsdb-1-4 root\]$ for h in \{4,5,1\}; do ssh opentsdb-1-$h /tony\_soft/zookeeper/bin/zkServer.sh start; done

\[hadoop@opentsdb-1-4 root\]$ for h in \{4,5,1\}; do ssh opentsdb-1-$h /tony\_soft/zookeeper/bin/zkServer.sh status; done \# 若成功启动则能看到follower或leader角色

**启动JN三个节点**：opentsdb-1-1、opentsdb-1-3、opentsdb-1-2

\[hadoop@opentsdb-1-4 ~\]$ for h in \{1,3,2\}; do ssh opentsdb-1-$h /tony\_soft/hadoop-2.8.4/sbin/hadoop-daemon.sh start journalnode; done

\[hadoop@opentsdb-1-4 ~\]$ callall jps \# 成功启动，在1、3、2机器上可看到JournalNode进程

在4号主机上仍然**以hadoop用户格式化HDFS**：

\[hadoop@opentsdb-1-4 ~\]$ hdfs namenode –format \# 若看到Exiting with status 0，则格式化成功

由于配了hdfs高可用，需要进行下一步：

**在4号机(nn1)启动第一个NN:**

\[hadoop@opentsdb-1-4 ~\]$ $HADOOP\_HOME/sbin/hadoop-daemon.sh start namenode

**在5号机(nn2)格式化HDFS:**

\[hadoop@opentsdb-1-4 ~\]$ ssh opentsdb-1-5 hdfs namenode –bootstrapStandby \# 若Exiting with status 0则格式化成功

**在nn1格式化zkfc:**

\[hadoop@opentsdb-1-4 ~\]$ hdfs zkfc –formatZK

**在nn1启动HDFS集群：**

\[hadoop@opentsdb-1-4 ~\]$ $HADOOP\_HOME/sbin/start-dfs.sh \# 验证结果: 和预期的一致

![20190505175320282.png][]

**在nn1启动YARN：**

\[hadoop@opentsdb-1-4 ~\]$ $HADOOP\_HOME/sbin/start-yarn.sh \# yarn配置了高可用,rm1和rm2要分别启动

**rm2****需要单独启动：**

\[hadoop@opentsdb-1-4 ~\]$ ssh opentsdb-1-5 $HADOOP\_HOME/sbin/yarn-daemon.sh start resourcemanager

![2019050517532095.png][]

原因是缺少libexec目录下的shell脚本,可在hdfs脚本找到DEFAULT\_LIBEXEC\_DIR目录，放到yarn脚本里：

![20190505175320103.png][]

再次尝试启动rm2：ssh opentsdb-1-5 $HADOOP\_HOME/sbin/yarn-daemon.sh start resourcemanager

![2019050517532092.png][]

\[hadoop@opentsdb-1-4 ~\]$ callall jps

正在opentsdb-1-4 执行命令jps

Warning: Permanently added 'opentsdb-1-4,192.168.35.2' (ECDSA) to the list of known hosts.

17873 Jps

14482 NameNode

16132 ResourceManager

13179 QuorumPeerMain

15646 DFSZKFailoverController

\--------命令执行成功！

正在opentsdb-1-5 执行命令jps

Warning: Permanently added 'opentsdb-1-5,192.168.35.3' (ECDSA) to the list of known hosts.

10053 NameNode

7961 QuorumPeerMain

10169 DFSZKFailoverController

10873 ResourceManager

11772 Jps

\--------命令执行成功！

正在opentsdb-1-1 执行命令jps

Warning: Permanently added 'opentsdb-1-1,192.168.35.4' (ECDSA) to the list of known hosts.

11168 Jps

9616 DataNode

7735 QuorumPeerMain

8491 JournalNode

10110 NodeManager

\--------命令执行成功！

正在opentsdb-1-3 执行命令jps

Warning: Permanently added 'opentsdb-1-3,192.168.35.5' (ECDSA) to the list of known hosts.

7717 JournalNode

9323 NodeManager

10379 Jps

8829 DataNode

\--------命令执行成功！

正在opentsdb-1-2 执行命令jps

Warning: Permanently added 'opentsdb-1-2,192.168.35.6' (ECDSA) to the list of known hosts.

8769 DataNode

9258 NodeManager

10315 Jps

7660 JournalNode

\--------命令执行成功

【配置并启动hbase集群】

hbase-env.sh 添加下列配置项

export JAVA\_HOME= /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.212.b04-0.el7\_6.x86\_64

export HBASE\_MANAGES\_ZK=false

\# Extra Java CLASSPATH elements. Optional.

export **HBASE\_CLASSPATH**=/tony\_soft/hadoop-2.8.4/etc/hadoop

hbase-site.xml高可用配置：

<name>hbase.rootdir</name>

<value>hdfs://tony/hbase</value>

<description>指定hdfs上的HRegionServers共享目录(hbase数据存放地)</description>

</property>

<name>hbase.cluster.distributed</name>

</property>

<name>hbase.tmp.dir</name>

<value>**/tony\_soft/hbase-1.4.9/logs**</value>

<description>指定缓存文件存储的路径</description>

</property>

<name>hbase.zookeeper.quorum</name>

<value>opentsdb-1-4:2181,opentsdb-1-5:2181,opentsdb-1-1:2181</value>

<description>指定ZooKeeper集群socket</description>

</property>

<name>hbase.zookeeper.property.dataDir</name>

<value>/zkdata</value>

<description>Zookeeper数据目录，与上述ZooKeeper集群相关配置一致</description>

</property>

</configuration>

regionservers

$HBASE\_HOME/conf/regionserver的内容和/tony\_soft/hadoop-2.8.4/etc/hadoop/slaves中的一致:

opentsdb-1-1

opentsdb-1-3

opentsdb-1-2

\# 分发配置文件到其他节点

\[root@opentsdb-1-4 conf\]\# send2others hbase-env.sh && send2others hbase-site.xml && send2others regionservers

启动HA-HBASE集群：

\[hadoop@opentsdb-1-4 conf\]$ /tony\_soft/hbase-1.4.9/bin/start-hbase.sh \# 启动失败，进一步查看日志

\[hadoop@opentsdb-1-4 hbase-1.4.9\]$ vim /tony\_soft/hbase-1.4.9/logs/hbase-hadoop-master-opentsdb-1-4.log

Caused by: java.lang.RuntimeException: Failed to create local dir /opt/hbase-2.0.0-alpha3/tmp/local/jars, DynamicClassLoader failed to init

排错1：修改hbase-env.sh，把hadoop的配置文件所在路径添加到HBASE\_CLASSPATH

export HBASE\_CLASSPATH=/tony\_soft/hadoop-2.8.4/etc/hadoop

\[root@opentsdb-1-4 conf\]\# send2others hbase-env.sh

再次启动，报错：

opentsdb-1-2: mkdir: cannot create directory ‘/tony\_soft/hbase-1.4.9/bin/../logs’: Permission denied

opentsdb-1-1: mkdir: cannot create directory ‘/tony\_soft/hbase-1.4.9/bin/../logs’: Permission denied

opentsdb-1-3: mkdir: cannot create directory ‘/tony\_soft/hbase-1.4.9/bin/../logs’: Permission denied

排错2：修复权限问题：

\[root@opentsdb-1-4 hbase-1.4.9\]\# callall chown -R hadoop:hadoop /tony\_soft/hbase-1.4.9 && callall chown -R hadoop:hadoop /tony\_soft/hbase/

\[hadoop@opentsdb-1-4 hbase-1.4.9\]$ for i in \{5,1,3,2\}; do ssh opentsdb-1-$i mkdir -p /tony\_soft/hbase-1.4.9/logs; done

\[hadoop@opentsdb-1-4 hbase-1.4.9\]$ callall ls -l /tony\_soft/hbase-1.4.9 | grep --color logs

\[hadoop@opentsdb-1-4 hbase-1.4.9\]$ /tony\_soft/hbase-1.4.9/bin/start-hbase.sh \# 若成功则本机能看到Hmaster服务，5号机没有Hmaster(因为需手动启动,不会自启)；另外能看到1、3、2号机上都有HregionServer服务

尝试在5号机上启动另一个Hmaster:

\[hadoop@opentsdb-1-4 conf\]$ ssh opentsdb-1-5 /tony\_soft/hbase/bin/hbase-daemon.sh start master \# 失败

启用高可用hbase：

先关闭hbase集群

\[hadoop@opentsdb-1-4 conf\]$ /tony\_soft/hbase-1.4.9/bin/stop-hbase.sh

再添加另一个主机名(或IP地址)到$HBASE\_HOME/conf/**backup-masters**并同步到其他节点

\[root@opentsdb-1-4 conf\]\# send2others backup-masters

\[root@opentsdb-1-4 conf\]\# callall chown hadoop:hadoop /tony\_soft/hbase-1.4.9/conf/backup-masters

再次启动hbase集群

\[hadoop@opentsdb-1-4 conf\]$ /tony\_soft/hbase-1.4.9/bin/start-hbase.sh \# 成功后可由jps看到相关进程

\[hadoop@opentsdb-1-4 hbase-1.4.9\]$ callall jps

正在opentsdb-1-4 执行命令jps

Warning: Permanently added 'opentsdb-1-4,192.168.35.2' (ECDSA) to the list of known hosts.

577 HMaster

14482 NameNode

16132 ResourceManager

13179 QuorumPeerMain

844 Jps

15646 DFSZKFailoverController

\--------命令执行成功！

正在opentsdb-1-5 执行命令jps

Warning: Permanently added 'opentsdb-1-5,192.168.35.3' (ECDSA) to the list of known hosts.

10053 NameNode

26182 Jps

7961 QuorumPeerMain

10169 DFSZKFailoverController

10873 ResourceManager

26012 HMaster

\--------命令执行成功！

正在opentsdb-1-1 执行命令jps

Warning: Permanently added 'opentsdb-1-1,192.168.35.4' (ECDSA) to the list of known hosts.

9616 DataNode

7735 QuorumPeerMain

23529 HRegionServer

23739 Jps

8491 JournalNode

10110 NodeManager

\--------命令执行成功！

正在opentsdb-1-3 执行命令jps

Warning: Permanently added 'opentsdb-1-3,192.168.35.5' (ECDSA) to the list of known hosts.

7717 JournalNode

22695 HRegionServer

22904 Jps

9323 NodeManager

8829 DataNode

\--------命令执行成功！

正在opentsdb-1-2 执行命令jps

Warning: Permanently added 'opentsdb-1-2,192.168.35.6' (ECDSA) to the list of known hosts.

8769 DataNode

9258 NodeManager

7660 JournalNode

22909 Jps

22686 HRegionServer

\--------命令执行成功！

### 【第三章】安装openTSDB ###

【须先安装Gnuplot】

为什么要安装gnuplot：Opentsdb依赖gnuplot 插件

\[root@opentsdb-1-4 conf\]\# callall yum install –y gnuplot \# 为所有节点安装

版本要求：openTSDB-2.3.0及以上，我这里下载一个2.4版的rpm包，以rpm方式安装OpenTSDB：

\[root@opentsdb-1-4 tony\_soft\]\# wget

[https://github.com/OpenTSDB/opentsdb/releases/download/v2.4.0/opentsdb-2.4.0.noarch.rpm][https_github.com_OpenTSDB_opentsdb_releases_download_v2.4.0_opentsdb-2.4.0.noarch.rpm]

**\#** **分布式安装openTSDB**

分发到其他节点send2others opentsdb-2.4.0.noarch.rpm

然后**每个节点安装**即可callall rpm –ivh opentsdb-2.4.0.noarch.rpm \# 默认会安装到 /usr/share/ 下

openTSDB本身没有分布式的实现方案，而是借助于HBase的分布式集群方案。也就是说openTSDB访问同一个HBase集群，返回相同的数据镜像。TSDMain是openTSDB的进程名。

vim /etc/opentsdb/opentsdb.conf

\[root@opentsdb-1-4 opentsdb\]\# egrep --color "^tsd" /etc/opentsdb/opentsdb.conf

tsd.network.port = 4040 \# 自定义HTTP端口号

tsd.http.staticroot = /usr/share/opentsdb/static/ \# 使用默认值即可

tsd.http.cachedir = /tmp/opentsdb \# 配置openTSDB的缓存目录

\# false改为true那么上传数据时会自动创建metric，否则会提示Unknown metric 的错误。也可设置为false但使用tsdb mkmetric proc.loadavg.1m来手动添加 metric

tsd.core.auto\_create\_metrics = true

tsd.core.plugin\_path = /usr/share/opentsdb/plugins \# 使用默认opentsDB的插件所在路径

tsd.storage.hbase.zk\_quorum = opentsdb-1-4,opentsdb-1-5,opentsdb-1-1 \# 配置zk三个主机名

\# 启动openTSdb之前，需要在Hbase中创建openTSDB所需的表

\[root@opentsdb-1-4 ~\]\# cd /usr/share/opentsdb/

\[root@opentsdb-1-4 opentsdb\]\# env COMPRESSION=NONE HBASE\_HOME=/tony\_soft/hbase-1.4.9 ./tools/**create\_table.sh** > /tmp/creatingForTbls4opentsDB.log 2>&1 &

\# 脚本会自动在default名称空间下创建4个表tsdb-uid,tsdb,tsdb-tree,tsdb-meta

\# 脚本运行失败，查看日志发现在创建default:tsdb表时有错误

create 'tsdb',

\{NAME => 't', VERSIONS => 1, COMPRESSION => 'NONE', BLOOMFILTER => 'ROW', DATA\_BLOCK\_ENCODING => 'DIFF', TTL => 'FOREVER'\}

ERROR: For input string: "FOREVER"

\# 排错：建表语句不被hbase认可(我安装的是hbase-1.4.9版本)，那么去掉FOREVER或者修改其值为2147483647

\[root@opentsdb-1-4 ~\]\# grep -C 3 --color "TSDB\_TTL" /usr/share/opentsdb/tools/create\_table.sh

\# This can save a lot of storage space.

DATA\_BLOCK\_ENCODING=$\{DATA\_BLOCK\_ENCODING-'DIFF'\}

DATA\_BLOCK\_ENCODING=\`echo "$DATA\_BLOCK\_ENCODING" | tr a-z A-Z\`

TSDB\_TTL=$\{TSDB\_TTL-'FOREVER'\}

case $COMPRESSION in

(NONE|LZO|GZIP|SNAPPY) :;; \# Known good.

\--

\{NAME => 'name', COMPRESSION => '$COMPRESSION', BLOOMFILTER => '$BLOOMFILTER', DATA\_BLOCK\_ENCODING => '$DATA\_BLOCK\_ENCODING'\}

create '$TSDB\_TABLE',

\{NAME => 't', VERSIONS => 1, COMPRESSION => '$COMPRESSION', BLOOMFILTER => '$BLOOMFILTER', DATA\_BLOCK\_ENCODING => '$DATA\_BLOCK\_ENCODING', TTL => '$TSDB\_TTL'\}

create '$TREE\_TABLE',

\{NAME => 't', VERSIONS => 1, COMPRESSION => '$COMPRESSION', BLOOMFILTER => '$BLOOMFILTER', DATA\_BLOCK\_ENCODING => '$DATA\_BLOCK\_ENCODING'\}

\# 排错之后重试

\[root@opentsdb-1-4 tools\]\# cd /usr/share/opentsdb/tools && vim create\_table.sh

\[root@opentsdb-1-4 tools\]\# grep -C 1 --color "TSDB\_TTL" create\_table.sh

DATA\_BLOCK\_ENCODING=\`echo "$DATA\_BLOCK\_ENCODING" | tr a-z A-Z\`

\#TSDB\_TTL=$\{TSDB\_TTL-'FOREVER'\}

TSDB\_TTL=$\{TSDB\_TTL-'2147483647'\}

\--

create '$TSDB\_TABLE',

\{NAME => 't', VERSIONS => 1, COMPRESSION => '$COMPRESSION', BLOOMFILTER => '$BLOOMFILTER', DATA\_BLOCK\_ENCODING => '$DATA\_BLOCK\_ENCODING', TTL => '$TSDB\_TTL'\}

\[root@opentsdb-1-4 tools\]\# send2others create\_table.sh

\[root@opentsdb-1-4 opentsdb\]\# pwd

/usr/share/opentsdb

\[root@opentsdb-1-4 opentsdb\]\# env COMPRESSION=NONE HBASE\_HOME=/tony\_soft/hbase-1.4.9 ./tools/create\_table.sh > /tmp/re\_creatingForTbls4opentsDB.log 2>&1 &

hbase(main):009:0> list **\# hbase shell****中可看到，成功建表**

TABLE

tsdb

tsdb-meta

tsdb-tree

tsdb-uid

4 row(s) in 0.0080 seconds

=> \["tsdb", "tsdb-meta", "tsdb-tree", "tsdb-uid"\]

\# 启动openTSDB集群(也可以仅启动其中1、2台)

\[root@opentsdb-1-4 bin\]\# callall "/usr/share/opentsdb/bin/tsdb tsd > /tmp/tsd\_running.log 2>&1 &"

\[root@opentsdb-1-4 bin\]\# callall "jps | grep -v grep | grep -i tsdmain" \# 可在所有节点看到TSDMain进程

\[root@opentsdb-1-4 bin\]\# callall "netstat -nltp | grep -i 4040" \# 可看到4、5、1、3、2号节点都有对应的4040端口号

\# 成功安装opentsdb后，使用客户端脚本向opentsdb插入示例数据

\[root@opentsdb-1-4 tmp\]\# cat openTSdb\_test.py

import requests

payload = \{

"metric": "sys.cpu.nice",

"timestamp": '1519544801',

"value": '29',

"tags": \{

"host": "web01",

"dc": "lga"

payload1 = \{

"metric": "sys.cpu.nice",

"timestamp": '1519544802',

"value": '30',

"tags": \{

"host": "web01",

"dc": "lga"

payload2 = \{

"metric": "sys.cpu.nice",

"timestamp": '1519544803',

"value": '29',

"tags": \{

"host": "web01",

"dc": "lga"

payload3 = \{

"metric": "sys.cpu.nice",

"timestamp": '1519544804',

"value": '30',

"tags": \{

"host": "web01",

"dc": "lga"

ls = \[payload, payload1, payload2, payload3\]

def send\_json(json):

\# 向主机opentsdb-1-5写入数据(因为该节点也有TSDMain)

r = requests.post("http://opentsdb-1-5:4242/api/put?details", json=json)

return r.text

def main():

print send\_json(ls)

if \_\_name\_\_ == "\_\_main\_\_":

main()

\# tsd返回成功，表明成功写入数据到hbase

\[root@opentsdb-1-4 tmp\]\# python openTSdb\_test.py

\{"success":4,"failed":0,"errors":\[\]\}

\# 到4、5、1其中一台主机(这三台有zk服务QuorumPeerMain)上的hbase shell检查是否已插入数据

\[root@opentsdb-1-4 ~\]\# ssh opentsdb-1-1

hbase(main):001:0> list

TABLE

tsdb

tsdb-meta

tsdb-tree

tsdb-uid

4 row(s) in 0.1660 seconds

=> \["tsdb", "tsdb-meta", "tsdb-tree", "tsdb-uid"\]

hbase(main):002:0> scan 'default:tsdb' \# 有数据

hbase(main):003:0> scan 'default:tsdb-meta' \# 没数据

hbase(main):004:0> scan 'default:tsdb-tree' \# 没数据

hbase(main):005:0> scan 'default:tsdb-uid' \# 有数据

以下是关于opentsdb的web restFul api安全使用的整理。这里是基于nginx做用户验证(也就是用户名+密码的方式)，是为了(云主机)内网用户在使用tsd的api过程中做个身份验证，只允许内网指定用户调用opentsdb的api。

### 【第四章】通过Nginx对OpenTSDB做用户验证，确保opentsdb http接口安全 ###

这种简单认证适用于公司内网，通过认证区分不同用户。

CentOS7安装Nginx可参考: [https://www.cnblogs.com/kaid/p/7640723.html][https_www.cnblogs.com_kaid_p_7640723.html]

Nginx 配置用户名密码登录可参考: [https://my.oschina.net/AmosWang/blog/2873864][https_my.oschina.net_AmosWang_blog_2873864]

【step 1】

openTSDB主要配置项：

![20190505175435237.png][]

批量启动TSDMain(即tsd服务)：callall "/usr/share/opentsdb/bin/tsdb tsd > /tmp/tsd\_running.log 2>&1 &"

【step 2】

安装Nginx过程省略。

安装htpasswd工具(生成用户名和密码并保存到文件)

在/usr/local/nginx/conf/nginx.conf的http代码块里的子块加入相关配置项：

server \{

\# 此处监听的端口号4242是Nginx对外暴露的端口(TSDMain的服务端口4040)

listen 4242;

auth\_basic "Password please";

\#密码地址

auth\_basic\_user\_file /usr/local/nginx/conf/ngpwd;

location / \{

\#代理地址：192.168.x.x是VPC内网,4040是OpenTSDB在内网的端口；外网需访问10.39.37.49:4242

proxy\_pass http://192.168.35.2:4040/;

proxy\_redirect default;

启动Nginx服务(略)。

在4号机使用htpasswd工具生成密码文件：

htpasswd -c /usr/local/soft/nginx/conf/pwd tony 接着输入两次密码即可

\#htpasswd -c：意思是创建一个加密文件

\#/usr/local/nginx/conf/ngpwd：表示加密文件的绝对路径(与上述nginx.conf中的auth\_basic\_user\_file一致)

\# tony：用户名（密码tonykidkid）

![20190505175435239.png][]

生成用户密码后，重启nginx服务使配置生效

![20190505175435235.png][]

验证用户认证功能：

（1）从外网浏览器访问[http://10.39.46.49:4242/][http_10.39.46.49_4242]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70][]

此时需要输入密码才能继续访问：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 1][]

（2）从外网采用curl [http://10.39.46.49:4242][http_10.39.46.49_4242 1] 未发送认证信息访问openTSDB时，HTTP返回码401：

![20190505175435278.png][]

验证成功。

从外网curl -u tony:tonykidkid [http://10.39.46.49:4242][http_10.39.46.49_4242 1] 发送用户认证信息访问openTSDB时，正常返回：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 2][]

验证成功。

（3）从VPC内网采用客户端脚本向openTSDB put数据，HTTP返回码401：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 3][]

验证成功。

（4）从内网访问,带上认证信息,正常返回：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 4][]

验证成功。

（5）使用客户端脚本向[http://192.168.35.2:4242/api/put?details=true][http_192.168.35.2_4242_api_put_details_true] 写入单点数据

python脚本Unauthed\_openTSdd.py源代码：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 5][]

未带认证信息时，返回401：

![20190505175435391.png][]

验证成功。

（6）使用客户端脚本以 [http://username:password@ip:port/api][http_username_password_ip_port_api] 方式传入认证信息

python脚本authed\_openTSdb\_test.py源码：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 6][]

认证通过，成功写入数据：

![20190505175435407.png][]

在hbase可查到：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 7][]

验证成功。

【全文结束】读者觉得有什么要补充的，或者哪里不明白的可以给我留言，觉得没问题的可以点个赞 ^\_^

[https_mirrors.tuna.tsinghua.edu.cn_apache_zookeeper_zookeeper-3.4.14_zookeeper-3.4.14.tar.gz]: https://mirrors.tuna.tsinghua.edu.cn/apache/zookeeper/zookeeper-3.4.14/zookeeper-3.4.14.tar.gz
[https_mirrors.tuna.tsinghua.edu.cn_apache_hadoop_common_hadoop-2.8.4_hadoop-2.8.4.tar.gz]: https://mirrors.tuna.tsinghua.edu.cn/apache/hadoop/common/hadoop-2.8.4/hadoop-2.8.4.tar.gz
[https_mirrors.tuna.tsinghua.edu.cn_apache_hbase_stable_hbase-1.4.9-bin.tar.gz]: https://mirrors.tuna.tsinghua.edu.cn/apache/hbase/stable/hbase-1.4.9-bin.tar.gz
[https_jingyan.baidu.com_article_bea41d43b558ffb4c51be6db.html]: https://jingyan.baidu.com/article/bea41d43b558ffb4c51be6db.html
[https_www.linuxidc.com_Linux_2015-11_124911.htm]: https://www.linuxidc.com/Linux/2015-11/124911.htm
[https_www.staroon.dev_2017_11_05_SetEnv]: https://www.staroon.dev/2017/11/05/SetEnv/
[2019050517532057.png]: /images/20220205/f09fdc55dfed4f5991234299cc032dd1.png
[20190505175320282.png]: /images/20220205/665980b2e854401d9708a2418ed59b1a.png
[2019050517532095.png]: /images/20220205/4e6b1710d45e4e3782f01f0dca1f8272.png
[20190505175320103.png]: /images/20220205/24aa5f1fe22a40e68d1af74040cc1518.png
[2019050517532092.png]: /images/20220205/e9df16731ae54e93ab4d170d7ab84943.png
[https_github.com_OpenTSDB_opentsdb_releases_download_v2.4.0_opentsdb-2.4.0.noarch.rpm]: https://github.com/OpenTSDB/opentsdb/releases/download/v2.4.0/opentsdb-2.4.0.noarch.rpm
[https_www.cnblogs.com_kaid_p_7640723.html]: https://www.cnblogs.com/kaid/p/7640723.html
[https_my.oschina.net_AmosWang_blog_2873864]: https://my.oschina.net/AmosWang/blog/2873864
[20190505175435237.png]: /images/20220205/a9e056e9acd7431c8d462bd3dbc55dd7.png
[20190505175435239.png]: /images/20220205/298425c95e0b4e52b3dfab4e8ce9241f.png
[20190505175435235.png]: /images/20220205/4c877dd7c9314b8eabc5e524f5c10925.png
[http_10.39.46.49_4242]: http://10.39.46.49:4242/
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70]: /images/20220205/b7b3165e9c7d477b8cbb736e75f5c622.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 1]: /images/20220205/6d9e106f37d04456892ed66468046b5a.png
[http_10.39.46.49_4242 1]: http://10.39.46.49:4242
[20190505175435278.png]: /images/20220205/4203ff91ded14f239f0e4868d279f671.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 2]: /images/20220205/f884781b3154414a842a80773491ba05.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 3]: /images/20220205/ba219ffab9184e4bb9d3a76043a9d4af.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 4]: /images/20220205/3a220c7b829c4887a122293b6abc1292.png
[http_192.168.35.2_4242_api_put_details_true]: http://192.168.25.2:4242/api/put?details=true
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 5]: /images/20220205/8885babd8c86437b94e5615e6d7d22c1.png
[20190505175435391.png]: /images/20220205/de278eff643f4bffa58dc6cfc8a23cab.png
[http_username_password_ip_port_api]: http://ip/api
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 6]: /images/20220205/2f8f84c96c874d9d942bd16ffc1995ba.png
[20190505175435407.png]: /images/20220205/f17116e50bc54744b7204d205c779c08.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzMxNTk4MTEz_size_16_color_FFFFFF_t_70 7]: /images/20220205/034f3b92d26c46ffa4576b4f04dbbcce.png