Saprk-日志实战
一、用户行为日志
1.概念
用户每次访问网站时所有的行为日志(访问、浏览、搜索、点击)
用户行为轨迹,流量日志
2.原因
分析日志:
网站页面访问量
网站的粘性
推荐
3.生产渠道
(1)Nginx
(2)Ajax
4.日志内容
日志数据内容:
1.访问的系统属性:操作系统、浏览器等
2.访问特征:点击URL,跳转页面(referer)、页面停留时间
3.访问信息:seesion_id、访问id信息(地市\运营商)
注意:Nginx配置,可以获取指定信息
5.意义
(1)网站的眼睛
投放广告收益
(2)网站的神经
网站布局(影响用户体验)
(3)网站的大脑
二、离线数据处理
1.处理流程
1)数据采集
Flume:
产生的Web日志,写入到HDFS
2)数据清洗
Spark\Hive\MapReduce--》HDFS(Hive/Spark SQL表)
3)数据处理
按照业务逻辑进行统计分析
Spark\Hive\MapReduce--》HDFS(Hive/Spark SQL表)
4)处理结果入库
RDBMS(MySQL)\NoSQL(HBase、Redis)
5)数据可视化展示
通过图形化展示:饼图、柱状图、地图、折线图
Echarts、HUE、Zeppelin
三、项目需求
code/video
需求一:
统计imooc主站最受欢迎的课程/手记Top N访问次数
需求二:
按地市统计imooc主站最受欢迎的Top N课程
a.根据IP地址获取出城市信息
b.窗口函数在Spark SQL中的使用
需求三:
按流量统计imooc主站最受欢迎的Top N课程
四、日志内容
需要字段:
访问时间、访问URL、访问过程耗费流量、访问IP地址
日志处理:
一般的日志处理方式,我们是需要进行分区的,
按照日志中的访问时间进行相应的分区,比如: d, h,m5(每5分钟一个分区)
输入:访问时间、访问URL、耗费的流量、访问IP地址信息
输出:URL、cmsType(video/article)、cmsId(编号)、流量、ip、城市信息、访问时间、天
Maven打包
mvn install:install-file -Dfile=D:\ipdatabase-master\ipdatabase-master\target\ipdatabase-1.0-SNAPSHOT.jar -DgroupId=com.ggstar2 -DartifactId=ipdatabase -Dversion=1.0 -Dpackaging=jar
五、数据清洗
1.原始日志解析
package com.saddam.spark.MuKe.ImoocProject
import org.apache.spark.SparkContext
import org.apache.spark.sql.SparkSession
/**
* 第一步清洗:抽取出所需要指定列数据
*
* 添加断点,可以查看各个字段
*/
object SparkStatFormatJob {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.appName("SparkStatFormatJob")
.master("local[2]")
.getOrCreate()
val logRDD = spark.sparkContext.textFile("D:\\Spark\\DataSets\\access.20161111.log")
// logRDD.take(10).foreach(println)
val result = logRDD.map(line => {
val split = line.split(" ")
val ip = split(0)
/**
* 原始日志的第三个和第四个字段拼接起来就是完整的时间字段:
* [10/Nov/2016:00:01:02 +0800]==>yyyy-MM-dd HH
*/
//TODO 使用时间解析工具类
val time = split(3) + " " + split(4)
//"http://www.imooc.com/code/1852" 引号需要放空
val url = split(11).replaceAll("\"", "")
val traffic = split(9)
//使用元组
// (ip,DateUtils.parse(time),url,traffic)
DateUtils.parse(time) + "\t" + url + "\t" + traffic + "\t" + ip
}).take(20).foreach(println)
// result.saveAsTextFile("D:\\Spark\\OutPut\\log_local_2")
/*
(10.100.0.1,[10/Nov/2016:00:01:02 +0800])
(117.35.88.11,[10/Nov/2016:00:01:02 +0800])
(182.106.215.93,[10/Nov/2016:00:01:02 +0800])
(10.100.0.1,[10/Nov/2016:00:01:02 +0800])
*/
spark.stop()
}
}
2.日期工具类
package com.saddam.spark.MuKe.ImoocProject
import java.util.{Date, Locale}
import org.apache.commons.lang3.time.FastDateFormat
/**
* 日期时间解析工具类
*/
object DateUtils {
// 输入文件日期时间格式
//[10/Nov/2016:00:01:02 +0800]
val YYYYMMDDHHMM_TIME_FOEMAT= FastDateFormat.getInstance("dd/MMM/yyyy:HH:mm:ss Z",Locale.ENGLISH)
//目标日期格式
//2016-11-10 00:01:02
val TARGET_FORMAT=FastDateFormat.getInstance("yyyy-MM-dd HH:mm:ss")
/**
*解析时间
* @param time
* @return
*/
def parse(time:String)={
TARGET_FORMAT.format(new Date(getTime(time)))
}
/**
* 获取输入日志时间:long类型
*
* time:[10/Nov/2016:00:01:02 +0800]
* @param time
* @return
*/
def getTime(time:String)= {
try {
YYYYMMDDHHMM_TIME_FOEMAT.parse(time.substring(time.indexOf("[") + 1, time.lastIndexOf("]"))).getTime
} catch {
case e: Exception => {
0l
}
}
}
def main(args: Array[String]): Unit = {
println(parse("[10/Nov/2016:00:01:02 +0800]"))
}
}
六、项目需求
需求一
统计imooc主站最受欢迎的课程/手记TopN访问次数
按照需求完成统计信息并将统计结果入库
--使用DataFrame API完成统计分析
--使用SQL API完成统计分析
package com.saddam.spark.MuKe
import java.sql.{Connection, DriverManager, PreparedStatement}
import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.sql.functions._
import scala.collection.mutable.ListBuffer
object PopularVideoVisits {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.appName("TopNStatJob")
.config("spark.sql.sources.partitionColumnTypeInference.enable", "false")
.master("local[2]")
.getOrCreate()
val accessDF=spark.read.format("parquet").load("D:\\Spark\\DataSets\\clean_city")
accessDF.printSchema()
accessDF.show(false)
/*
+----------------------------------+-------+-----+-------+---------------+----+-------------------+--------+
|url |cmsType|cmsId|traffic|ip |city|time |day |
+----------------------------------+-------+-----+-------+---------------+----+-------------------+--------+
|http://www.imooc.com/video/4500 |video |4500 |304 |218.75.35.226 |鏈煡 |2017-05-11 14:09:14|20170511|
|http://www.imooc.com/video/14623 |video |14623|69 |202.96.134.133 |鏈煡 |2017-05-11 15:25:05|20170511|
|http://www.imooc.com/article/17894|article|17894|115 |202.96.134.133 |鏈煡 |2017-05-11 07:50:01|20170511|
*/
//代码重构
val day="20170511"
videoAccessTopNStat(spark,accessDF,day)
//MySQL工具类测试
println(MySQLUtils.getConnection())
/**
* 按照流量进行统计
*/
def videoAccessTopNStat(spark:SparkSession,accessDF:DataFrame,day:String)={
//隐式转换
import spark.implicits._
//TODO 统计方式一:DataFrame方式统计video
val videoAccessTopNDF= accessDF
.filter($"day"===day && $"cmsType"==="video")
.groupBy("day","cmsId")
.agg(count("cmsId").as("times"))
videoAccessTopNDF.printSchema()
videoAccessTopNDF.show(false)
//TODO 统计方式二:SQL方式统计article
accessDF.createOrReplaceTempView("temp")
val videoAccessTopNSQL = spark.sql("select " +
"day,cmsId,count(1) as times " +
"from temp " +
"where day='20170511' and cmsType='article' " +
"group by day,cmsId " +
"order by times desc")
videoAccessTopNSQL.show(false)
/**
* TODO 将最受欢迎的TopN课程统计结果写入MySQL
*
*/
try{
videoAccessTopNSQL.foreachPartition(partitionOfRecords=>{
val list =new ListBuffer[DayVideoAccessStat]
partitionOfRecords.foreach(info=>{
val day=info.getAs[Integer]("day").toString
val cmsId=info.getAs[Long]("cmsId")
val times=info.getAs[Long]("times")
list.append(DayVideoAccessStat(day,cmsId,times))
})
StatDAO.insertDayVideoAccessTopN(list)
})}catch {
case e:Exception=>e.printStackTrace()
}
}
spark.stop()
}
/**
* 课程访问次数实体类
*/
case class DayVideoAccessStat(day:String,cmsId:Long,times:Long)
/**
* TODO MySQL操作工具类
*/
object MySQLUtils{
def getConnection()={
DriverManager.getConnection("jdbc:mysql://121.37.2x.xx:3306/imooc_project?user=root&password=xxxxxx&useSSL=false")
}
/**
* 释放数据库连接等资源
* @param connection
* @param pstmt
*/
def release(connection: Connection, pstmt: PreparedStatement): Unit = {
try {
if (pstmt != null) {
pstmt.close()
}
} catch {
case e: Exception => e.printStackTrace()
} finally {
if (connection != null) {
connection.close()
}
}
}
}
/**
* TODO DAO数据库接口
*/
object StatDAO{
/**
* 批量保存DayVideoAccessStat到数据库
* insertDayVideoAccessTopN:每天访问视频的
*/
def insertDayVideoAccessTopN(list: ListBuffer[DayVideoAccessStat]): Unit = {
var connection:Connection = null
var pstmt:PreparedStatement = null
try {
connection =MySQLUtils.getConnection()
connection.setAutoCommit(false) //设置手动提交
val sql = "insert into day_video_access_topn_stat2(day,cms_id,times) values (?,?,?)"
pstmt = connection.prepareStatement(sql)
for (ele <- list) {
pstmt.setString(1, ele.day)
pstmt.setLong(2, ele.cmsId)
pstmt.setLong(3, ele.times)
pstmt.addBatch()
}
pstmt.executeBatch() // 执行批量处理
connection.commit() //手工提交
} catch {
case e: Exception => e.printStackTrace()
} finally {
MySQLUtils.release(connection, pstmt)
}
}
}
}
需求二
按地市统计imooc主站最受欢迎的Top N课程
package com.saddam.spark.MuKe
import java.sql.{Connection, DriverManager, PreparedStatement}
import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.sql.functions._
import scala.collection.mutable.ListBuffer
object PopularCiytVideoVisits {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.appName("TopNStatJob")
.config("spark.sql.sources.partitionColumnTypeInference.enable", "false")
.master("local[2]")
.getOrCreate()
val accessDF=spark.read.format("parquet").load("D:\\Spark\\DataSets\\clean_city")
accessDF.printSchema()
accessDF.show(false)
//代码重构
val day="20170511"
//TODO 按照地市进行统计TopN课程
cityAccessTopNStat(spark,accessDF,day)
/**
* 按照地市进行统计TopN课程
* @param spark
* @param accessDf
*/
def cityAccessTopNStat(spark: SparkSession,accessDF:DataFrame,day:String)={
import spark.implicits._
val cityAccessTopNDF=accessDF.filter($"day"===day&&$"cmsType"==="video").groupBy("day","city","cmsId").agg(count("cmsId").as("times")).orderBy($"times".desc)
cityAccessTopNDF.printSchema()
cityAccessTopNDF.show(false)
//Windows函数在Spark SQL的使用
val top3DF=cityAccessTopNDF.select(
cityAccessTopNDF("day"),
cityAccessTopNDF("city"),
cityAccessTopNDF("cmsId"),
cityAccessTopNDF("times"),
row_number().over(Window.partitionBy(cityAccessTopNDF("city"))
.orderBy(cityAccessTopNDF("times").desc)
).as("times_rank")
).filter("times_rank <=3") //.show(false) //Top3
/**
* 将地市进行统计TopN课程统计结果写入MySQL
*
*/
try{
top3DF.foreachPartition(partitionOfRecords=>{
val list =new ListBuffer[DayCityVideoAccessStat]
partitionOfRecords.foreach(info=>{
val day=info.getAs[Integer]("day").toString
val cmsId=info.getAs[Long]("cmsId")
val city=info.getAs[String]("city")
val times=info.getAs[Long]("times")
val timesRank=info.getAs[Int]("times_rank")
list.append(DayCityVideoAccessStat(day,cmsId,city,times,timesRank))
})
StatDAO.insertDayCityVideoAccessTopN(list)
})}catch {
case e:Exception=>e.printStackTrace()
}
}
spark.stop()
}
/**
* 实体类
*/
case class DayCityVideoAccessStat(day:String, cmsId:Long, city:String,times:Long,timesRank:Int)
/**
* TODO MySQL操作工具类
*/
object MySQLUtils{
def getConnection()={
DriverManager.getConnection("jdbc:mysql://121.37.2x.xx:3306/imooc_project?user=root&password=xxxxxx&useSSL=false")
}
/**
* 释放数据库连接等资源
* @param connection
* @param pstmt
*/
def release(connection: Connection, pstmt: PreparedStatement): Unit = {
try {
if (pstmt != null) {
pstmt.close()
}
} catch {
case e: Exception => e.printStackTrace()
} finally {
if (connection != null) {
connection.close()
}
}
}
}
/**
* TODO DAO数据库接口
*/
object StatDAO{
/**
* 批量保存DayCityVideoAccessStat到数据库
*/
def insertDayCityVideoAccessTopN(list: ListBuffer[DayCityVideoAccessStat]): Unit = {
var connection: Connection = null
var pstmt: PreparedStatement = null
try {
connection = MySQLUtils.getConnection()
connection.setAutoCommit(false) //设置手动提交
val sql = "insert into day_video_city_access_topn_stat(day,cms_id,city,times,times_rank) values (?,?,?,?,?) "
pstmt = connection.prepareStatement(sql)
for (ele <- list) {
pstmt.setString(1, ele.day)
pstmt.setLong(2, ele.cmsId)
pstmt.setString(3, ele.city)
pstmt.setLong(4, ele.times)
pstmt.setInt(5, ele.timesRank)
pstmt.addBatch()
}
pstmt.executeBatch() // 执行批量处理
connection.commit() //手工提交
} catch {
case e: Exception => e.printStackTrace()
} finally {
MySQLUtils.release(connection, pstmt)
}
}
}
}
需求三
按流量统计imooc主站最受欢迎的Top N课程
package com.saddam.spark.MuKe
import java.sql.{Connection, DriverManager, PreparedStatement}
import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.sql.functions._
import scala.collection.mutable.ListBuffer
object VideoTrafficVisits {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.appName("TopNStatJob")
.config("spark.sql.sources.partitionColumnTypeInference.enable", "false")
.master("local[2]")
.getOrCreate()
val accessDF=spark.read.format("parquet").load("D:\\Spark\\DataSets\\clean_city")
accessDF.printSchema()
accessDF.show(false)
//代码重构
val day="20170511"
//TODO 按照流量进行统计
videoTrafficsTopNStat(spark,accessDF,day)
/**
* 按照流量进行统计
*/
def videoTrafficsTopNStat(spark: SparkSession, accessDF:DataFrame,day:String): Unit = {
import spark.implicits._
val cityAccessTopNDF = accessDF.filter($"day" === day && $"cmsType" === "video")
.groupBy("day", "cmsId").agg(sum("traffic").as("traffics"))
.orderBy($"traffics".desc)
//.show(false)
/**
* 将流量进行统计TopN课程统计结果写入MySQL
*
*/
try{
cityAccessTopNDF.foreachPartition(partitionOfRecords=>{
val list =new ListBuffer[DayVideoTrafficsStat]
partitionOfRecords.foreach(info=>{
val day=info.getAs[Integer]("day").toString
val cmsId=info.getAs[Long]("cmsId")
val traffics=info.getAs[Long]("traffics")
list.append(DayVideoTrafficsStat(day,cmsId,traffics))
})
StatDAO.insertDayVideoTrafficsAccessTopN(list)
})}catch {
case e:Exception=>e.printStackTrace()
}
}
spark.stop()
}
/**
* 实体类
*/
case class DayVideoTrafficsStat(day:String,cmsId:Long,traffics:Long)
/**
* TODO MySQL操作工具类
*/
object MySQLUtils{
def getConnection()={
DriverManager.getConnection("jdbc:mysql://121.37.2x.2xx1:3306/imooc_project?user=root&password=xxxxxx&useSSL=false")
}
/**
* 释放数据库连接等资源
* @param connection
* @param pstmt
*/
def release(connection: Connection, pstmt: PreparedStatement): Unit = {
try {
if (pstmt != null) {
pstmt.close()
}
} catch {
case e: Exception => e.printStackTrace()
} finally {
if (connection != null) {
connection.close()
}
}
}
}
/**
* TODO DAO数据库接口
*/
object StatDAO{
/**
* 批量保存DayVideoTrafficsStat到数据库
*/
def insertDayVideoTrafficsAccessTopN(list: ListBuffer[DayVideoTrafficsStat]): Unit = {
var connection: Connection = null
var pstmt: PreparedStatement = null
try {
connection = MySQLUtils.getConnection()
connection.setAutoCommit(false) //设置手动提交
val sql = "insert into day_video_traffics_topn_stat(day,cms_id,traffics) values (?,?,?) "
pstmt = connection.prepareStatement(sql)
for (ele <- list) {
pstmt.setString(1, ele.day)
pstmt.setLong(2, ele.cmsId)
pstmt.setLong(3, ele.traffics)
pstmt.addBatch()
}
pstmt.executeBatch() // 执行批量处理
connection.commit() //手工提交
} catch {
case e: Exception => e.printStackTrace()
} finally {
MySQLUtils.release(connection, pstmt)
}
}
}
}
删除已有数据
/**
* 删除指定日期的数据
*/
def deleteData(day: String): Unit = {
val tables = Array("day_video_access_topn_stat",
"day_video_city_access_topn_stat",
"day_video_traffics_topn_stat")
var connection:Connection = null
var pstmt:PreparedStatement = null
try{
connection = MySQLUtils.getConnection()
for(table <- tables) {
// delete from table ....
val deleteSQL = s"delete from $table where day = ?"
pstmt = connection.prepareStatement(deleteSQL)
pstmt.setString(1, day)
pstmt.executeUpdate()
}
}catch {
case e:Exception => e.printStackTrace()
} finally {
MySQLUtils.release(connection, pstmt)
}
}
七、Zeppelin
官网
https://zeppelin.apache.org/
1.解压缩
[root@hadoop src]# tar zxvf zeppelin-0.7.1-bin-all
2.改名
[root@hadoop src]# mv zeppelin-0.7.1-bin-all zeppelin
3.启动
[root@hadoop bin]# ./zeppelin-daemon.sh start
4.Web界面
http://121.37.2xx.xx:8080
5.修改JDBC驱动
com.mysql.jdbc.Driver
xxxxxx
jdbc:mysql://121.37.2x.xx:3306/imooc_project?
root
#mysql驱动
/usr/local/src/mysql-connector-java-5.1.27-bin.jar
6.创建note
7.查询表
%jdbc
show tables;
8.图形展示
%jdbc
select cms_id,times from day_video_access_topn_stat;
八、Spark on Yarn
Spark运行模式
1)Local:开发时使用
2)Standalone:Spark自带的,若一个集群是standalone,则需要在多台机器上同时部署Spark
3)YARN:建议生产上使用该模式,统一使用yarn进行整个集群作业(MR、Spark)的资源调度
4)Mesos
不管使用那种模式,Spark应用程序代码是一模一样的,只需要在提交的时候指定--master指定
1.概述
Spark支持可插拔的集群管理模式
对于yarn而言,Spark Application仅仅只是一个客户端而已
2.client模式
Driver运行在Client端(提交Spark作业的机器)
Client会和请求到的Container进行通信来完成作业的调度和执行,Client是不能退出的
日志信息在控制台输出,便于我们测试
3.cluster模式
Driver运行在ApplicationMaster中
Client提交完作业就可以关掉,因为作业已在
Yarn上运行了
日志在终端输出,看控制台不到的,因为日志在Driver端,只能通过yarn logs -applicationId
4.两种模式对比
Driver运行位置
ApplicationMaster的职责
运行输出日志的位置
5.案例
设置HADOOP_CONF_DIR=?
Client模式
./bin/spark-submit \
--class org.apache.spark.examples.SparkPi \
--master yarn \
--executor-memory 1G \
--num-executors 1 \
/usr/local/src/spark/examples/jars/spark-examples_2.11-2.1.1.jar \
4
Pi is roughly 3.1411378528446323
22/02/28 18:52:26 INFO server.ServerConnector: Stopped Spark@1b0a7baf{HTTP/1.1}{0.0.0.0:4040}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@8a589a2{/stages/stage/kill,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@192f2f27{/jobs/job/kill,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@1bdf8190{/api,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@4f8969b0{/,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6fefce9e{/static,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@74cec793{/executors/threadDump/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@f9b7332{/executors/threadDump,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@18e7143f{/executors/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@209775a9{/executors,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@5db4c359{/environment/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@2c177f9e{/environment,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@33617539{/storage/rdd/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@47874b25{/storage/rdd,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@290b1b2e{/storage/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@1fc0053e{/storage,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@77307458{/stages/pool/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@389adf1d{/stages/pool,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@7bf9b098{/stages/stage/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@72e34f77{/stages/stage,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6e9319f{/stages/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6fa590ba{/stages,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@2416a51{/jobs/job/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@293bb8a5{/jobs/job,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@37ebc9d8{/jobs/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@5217f3d0{/jobs,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO ui.SparkUI: Stopped Spark web UI at http://192.168.184.135:4040
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Interrupting monitor thread
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Shutting down all executors
22/02/28 18:52:26 INFO cluster.YarnSchedulerBackend$YarnDriverEndpoint: Asking each executor to shut down
22/02/28 18:52:26 INFO cluster.SchedulerExtensionServices: Stopping SchedulerExtensionServices
(serviceOption=None,
services=List(),
started=false)
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Stopped
22/02/28 18:52:26 INFO spark.MapOutputTrackerMasterEndpoint: MapOutputTrackerMasterEndpoint stopped!
22/02/28 18:52:26 INFO memory.MemoryStore: MemoryStore cleared
22/02/28 18:52:26 INFO storage.BlockManager: BlockManager stopped
22/02/28 18:52:26 INFO storage.BlockManagerMaster: BlockManagerMaster stopped
22/02/28 18:52:26 INFO scheduler.OutputCommitCoordinator$OutputCommitCoordinatorEndpoint: OutputCommitCoordinator stopped!
22/02/28 18:52:26 INFO spark.SparkContext: Successfully stopped SparkContext
22/02/28 18:52:26 INFO util.ShutdownHookManager: Shutdown hook called
22/02/28 18:52:26 INFO util.ShutdownHookManager: Deleting directory /tmp/spark-a95834c0-d38b-457b-89b2-fed00d5bef56
Cluster模式
./bin/spark-submit \
--class org.apache.spark.examples.SparkPi \
--master yarn-cluster \
--executor-memory 1G \
--num-executors 1 \
/usr/local/src/spark/examples/jars/spark-examples_2.11-2.1.1.jar \
4
Pi is roughly 3.1411378528446323
22/02/28 18:52:26 INFO server.ServerConnector: Stopped Spark@1b0a7baf{HTTP/1.1}{0.0.0.0:4040}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@8a589a2{/stages/stage/kill,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@192f2f27{/jobs/job/kill,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@1bdf8190{/api,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@4f8969b0{/,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6fefce9e{/static,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@74cec793{/executors/threadDump/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@f9b7332{/executors/threadDump,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@18e7143f{/executors/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@209775a9{/executors,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@5db4c359{/environment/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@2c177f9e{/environment,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@33617539{/storage/rdd/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@47874b25{/storage/rdd,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@290b1b2e{/storage/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@1fc0053e{/storage,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@77307458{/stages/pool/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@389adf1d{/stages/pool,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@7bf9b098{/stages/stage/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@72e34f77{/stages/stage,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6e9319f{/stages/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@6fa590ba{/stages,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@2416a51{/jobs/job/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@293bb8a5{/jobs/job,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@37ebc9d8{/jobs/json,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO handler.ContextHandler: Stopped o.s.j.s.ServletContextHandler@5217f3d0{/jobs,null,UNAVAILABLE,@Spark}
22/02/28 18:52:26 INFO ui.SparkUI: Stopped Spark web UI at http://192.168.184.135:4040
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Interrupting monitor thread
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Shutting down all executors
22/02/28 18:52:26 INFO cluster.YarnSchedulerBackend$YarnDriverEndpoint: Asking each executor to shut down
22/02/28 18:52:26 INFO cluster.SchedulerExtensionServices: Stopping SchedulerExtensionServices
(serviceOption=None,
services=List(),
started=false)
22/02/28 18:52:26 INFO cluster.YarnClientSchedulerBackend: Stopped
22/02/28 18:52:26 INFO spark.MapOutputTrackerMasterEndpoint: MapOutputTrackerMasterEndpoint stopped!
22/02/28 18:52:26 INFO memory.MemoryStore: MemoryStore cleared
22/02/28 18:52:26 INFO storage.BlockManager: BlockManager stopped
22/02/28 18:52:26 INFO storage.BlockManagerMaster: BlockManagerMaster stopped
22/02/28 18:52:26 INFO scheduler.OutputCommitCoordinator$OutputCommitCoordinatorEndpoint: OutputCommitCoordinator stopped!
22/02/28 18:52:26 INFO spark.SparkContext: Successfully stopped SparkContext
22/02/28 18:52:26 INFO util.ShutdownHookManager: Shutdown hook called
22/02/28 18:52:26 INFO util.ShutdownHookManager: Deleting directory /tmp/spark-a95834c0-d38b-457b-89b2-fed00d5bef56
[root@hadoop01 spark]# ./bin/spark-submit --class org.apache.spark.examples.SparkPi --master yarn-cluster --executor-memory 1G --num-executors 1 /usr/local/src/spark/examples/jars/spark-examples_2.11-2.1.1.jar 4
Warning: Master yarn-cluster is deprecated since 2.0. Please use master "yarn" with specified deploy mode instead.
22/02/28 18:54:32 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
22/02/28 18:54:32 WARN util.Utils: Your hostname, hadoop01.localdomain resolves to a loopback address: 127.0.0.1; using 192.168.184.135 instead (on interface ens33)
22/02/28 18:54:32 WARN util.Utils: Set SPARK_LOCAL_IP if you need to bind to another address
22/02/28 18:54:32 INFO client.RMProxy: Connecting to ResourceManager at hadoop01/192.168.184.135:8032
22/02/28 18:54:32 INFO yarn.Client: Requesting a new application from cluster with 1 NodeManagers
22/02/28 18:54:33 INFO yarn.Client: Verifying our application has not requested more than the maximum memory capability of the cluster (8192 MB per container)
22/02/28 18:54:33 INFO yarn.Client: Will allocate AM container, with 1408 MB memory including 384 MB overhead
22/02/28 18:54:33 INFO yarn.Client: Setting up container launch context for our AM
22/02/28 18:54:33 INFO yarn.Client: Setting up the launch environment for our AM container
22/02/28 18:54:33 INFO yarn.Client: Preparing resources for our AM container
22/02/28 18:54:33 WARN yarn.Client: Neither spark.yarn.jars nor spark.yarn.archive is set, falling back to uploading libraries under SPARK_HOME.
22/02/28 18:54:35 INFO yarn.Client: Uploading resource file:/tmp/spark-c7e3fb91-c7d0-4f59-86ba-705a9f256144/__spark_libs__3085975169933820625.zip -> hdfs://hadoop01:9000/user/root/.sparkStaging/application_1646041633964_0004/__spark_libs__3085975169933820625.zip
22/02/28 18:54:35 INFO yarn.Client: Uploading resource file:/usr/local/src/spark/examples/jars/spark-examples_2.11-2.1.1.jar -> hdfs://hadoop01:9000/user/root/.sparkStaging/application_1646041633964_0004/spark-examples_2.11-2.1.1.jar
22/02/28 18:54:35 INFO yarn.Client: Uploading resource file:/tmp/spark-c7e3fb91-c7d0-4f59-86ba-705a9f256144/__spark_conf__2818552262823480245.zip -> hdfs://hadoop01:9000/user/root/.sparkStaging/application_1646041633964_0004/__spark_conf__.zip
22/02/28 18:54:35 INFO spark.SecurityManager: Changing view acls to: root
22/02/28 18:54:35 INFO spark.SecurityManager: Changing modify acls to: root
22/02/28 18:54:35 INFO spark.SecurityManager: Changing view acls groups to:
22/02/28 18:54:35 INFO spark.SecurityManager: Changing modify acls groups to:
22/02/28 18:54:35 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(root); groups with view permissions: Set(); users with modify permissions: Set(root); groups with modify permissions: Set()
22/02/28 18:54:35 INFO yarn.Client: Submitting application application_1646041633964_0004 to ResourceManager
22/02/28 18:54:35 INFO impl.YarnClientImpl: Submitted application application_1646041633964_0004
22/02/28 18:54:36 INFO yarn.Client: Application report for application_1646041633964_0004 (state: ACCEPTED)
22/02/28 18:54:36 INFO yarn.Client:
client token: N/A
diagnostics: N/A
ApplicationMaster host: N/A
ApplicationMaster RPC port: -1
queue: default
start time: 1646045675928
final status: UNDEFINED
tracking URL: http://hadoop01:8088/proxy/application_1646041633964_0004/
user: root
22/02/28 18:54:37 INFO yarn.Client: Application report for application_1646041633964_0004 (state: ACCEPTED)
22/02/28 18:54:38 INFO yarn.Client: Application report for application_1646041633964_0004 (state: ACCEPTED)
22/02/28 18:54:39 INFO yarn.Client: Application report for application_1646041633964_0004 (state: RUNNING)
22/02/28 18:54:39 INFO yarn.Client:
client token: N/A
diagnostics: N/A
ApplicationMaster host: 192.168.184.135
ApplicationMaster RPC port: 0
queue: default
start time: 1646045675928
final status: UNDEFINED
tracking URL: http://hadoop01:8088/proxy/application_1646041633964_0004/
user: root
22/02/28 18:54:40 INFO yarn.Client: Application report for application_1646041633964_0004 (state: RUNNING)
22/02/28 18:54:41 INFO yarn.Client: Application report for application_1646041633964_0004 (state: RUNNING)
22/02/28 18:54:42 INFO yarn.Client: Application report for application_1646041633964_0004 (state: RUNNING)
22/02/28 18:54:43 INFO yarn.Client: Application report for application_1646041633964_0004 (state: FINISHED)
22/02/28 18:54:43 INFO yarn.Client:
client token: N/A
diagnostics: N/A
ApplicationMaster host: 192.168.184.135
ApplicationMaster RPC port: 0
queue: default
start time: 1646045675928
final status: SUCCEEDED
tracking URL: http://hadoop01:8088/proxy/application_1646041633964_0004/
user: root
22/02/28 18:54:43 INFO util.ShutdownHookManager: Shutdown hook called
22/02/28 18:54:43 INFO util.ShutdownHookManager: Deleting directory /tmp/spark-c7e3fb91-c7d0-4f59-86ba-705a9f256144
6.检测ID
[root@hadoop01 spark]# yarn logs -applicationId application_1646041633964_0003
22/02/28 18:59:05 INFO client.RMProxy: Connecting to ResourceManager at hadoop01/192.168.184.135:8032
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/usr/local/src/hadoop/share/hadoop/common/lib/slf4j-log4j12-1.7.10.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/usr/local/src/hive/lib/log4j-slf4j-impl-2.6.2.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
/tmp/logs/root/logs/application_1646041633964_0003 does not exist.
Log aggregation has not completed or is not enabled.
#未指定参数,看不到,未作聚合日志配置,需要通过webUI页面
7.WebUI查看结果
http://hadoop01:8042/node/containerlogs/container_1646041633964_0004_01_000001/root
九、Spark项目运行到YARN
maven打包依赖
1.IDEA项目代码-词频统计
package com.bigdata
import org.apache.spark.sql.SparkSession
object WordCountYARN {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.getOrCreate()
if(args.length!=2){
println("Usage:WordCountYARN <inputPath><outputPath>")
}
val Array(inputPath,outputPath)=args
val rdd = spark.sparkContext.textFile(inputPath)
val df = rdd.flatMap(x=>x.split("\t")).map(word=>(word,1)).reduceByKey((a,b)=>(a+b))
df.saveAsTextFile(outputPath)
spark.stop()
}
}
2.spark-submit
spark-submit \
--class com.bigdata.WordCountYARN \
--name WordCount \
--master yarn \
--executor-memory 1G \
--num-executors 1 \
/usr/local/src/spark/spark_jar/BYGJ.jar \
hdfs://hadoop01:9000/wordcount.txt hdfs://hadoop01:9000/wc_output
3.查询结果
[root@hadoop01 spark_jar]# hadoop fs -cat /wc_output/part-*
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/usr/local/src/hadoop/share/hadoop/common/lib/slf4j-log4j12-1.7.10.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/usr/local/src/hive/lib/log4j-slf4j-impl-2.6.2.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
(hive,5)
(spark,5)
(hadoop,2)
(hbase,3)
-----------------------------------------
1.IDEA项目代码-日志清洗
package com.saddam.spark.MuKe.ImoocProject.LogClean
import org.apache.spark.sql.{SaveMode, SparkSession}
object SparkStatCleanJobYarn {
def main(args: Array[String]): Unit = {
val spark=SparkSession
.builder()
.getOrCreate()
if(args.length!=2){
println("Usage:WordCountYARN <inputPath><outputPath>")
}
val Array(inputPath,outputPath)=args
val accessRDD = spark.sparkContext.textFile(inputPath)
//TODO RDD->DF
val accessDF=spark.createDataFrame(accessRDD.map(x=>AccessConvertUtil.parseLog(x)),AccessConvertUtil.struct)
accessDF
.coalesce(1)
.write
.format("parquet")
.mode(SaveMode.Overwrite)
.partitionBy("day")
.save(outputPath)
spark.stop()
}
}
2.spark-submit
spark-submit \
--class com.saddam.spark.MuKe.ImoocProject.LogClean.SparkStatCleanJobYarn \
--name SparkStatCleanJobYarn \
--master yarn \
--executor-memory 1G \
--num-executors 1 \
--files /usr/local/src/spark/spark_jar/ipDatabase.csv,/usr/local/src/spark/spark_jar/ipRegion.xlsx \
/usr/local/src/spark/spark_jar/Spark.jar \
hdfs://hadoop01:9000/access.log hdfs://hadoop01:9000/log_output
3.查询结果
进入spark-shell
[root@hadoop01 datas]# spark-shell --master local[2] --jars /usr/local/src/mysql-connector-java-5.1.27-bin.jar
获取hdfs输出文件
/log_output/day=20170511/part-00000-36e30abb-3e42-4237-ad9f-a9f93258d4b2.snappy.parquet
读取文件
scala> spark.read.format("parquet").parquet("/log_output/day=20170511/part-00000-36e30abb-3e42-4237-ad9f-a9f93258d4b2.snappy.parquet").show(false)
SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
SLF4J: Defaulting to no-operation (NOP) logger implementation
SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details.
+----------------------------------+-------+-----+-------+---------------+----+-------------------+
|url |cmsType|cmsId|traffic|ip |city|time |
+----------------------------------+-------+-----+-------+---------------+----+-------------------+
|http://www.imooc.com/video/4500 |video |4500 |304 |218.75.35.226 | |2017-05-11 14:09:14|
|http://www.imooc.com/video/14623 |video |14623|69 |202.96.134.133 | |2017-05-11 15:25:05|
|http://www.imooc.com/article/17894|article|17894|115 |202.96.134.133 | |2017-05-11 07:50:01|
|http://www.imooc.com/article/17896|article|17896|804 |218.75.35.226 | |2017-05-11 02:46:43|
|http://www.imooc.com/article/17893|article|17893|893 |222.129.235.182| |2017-05-11 09:30:25|
|http://www.imooc.com/article/17891|article|17891|407 |218.75.35.226 | |2017-05-11 08:07:35|
|http://www.imooc.com/article/17897|article|17897|78 |202.96.134.133 | |2017-05-11 19:08:13|
|http://www.imooc.com/article/17894|article|17894|658 |222.129.235.182| |2017-05-11 04:18:47|
|http://www.imooc.com/article/17893|article|17893|161 |58.32.19.255 | |2017-05-11 01:25:21|
|http://www.imooc.com/article/17895|article|17895|701 |218.22.9.56 | |2017-05-11 13:37:22|
|http://www.imooc.com/article/17892|article|17892|986 |218.75.35.226 | |2017-05-11 05:53:47|
|http://www.imooc.com/video/14540 |video |14540|987 |58.32.19.255 | |2017-05-11 18:44:56|
|http://www.imooc.com/article/17892|article|17892|610 |218.75.35.226 | |2017-05-11 17:48:51|
|http://www.imooc.com/article/17893|article|17893|0 |218.22.9.56 | |2017-05-11 16:20:03|
|http://www.imooc.com/article/17891|article|17891|262 |58.32.19.255 | |2017-05-11 00:38:01|
|http://www.imooc.com/video/4600 |video |4600 |465 |218.75.35.226 | |2017-05-11 17:38:16|
|http://www.imooc.com/video/4600 |video |4600 |833 |222.129.235.182| |2017-05-11 07:11:36|
|http://www.imooc.com/article/17895|article|17895|320 |222.129.235.182| |2017-05-11 19:25:04|
|http://www.imooc.com/article/17898|article|17898|460 |202.96.134.133 | |2017-05-11 15:14:28|
|http://www.imooc.com/article/17899|article|17899|389 |222.129.235.182| |2017-05-11 02:43:15|
+----------------------------------+-------+-----+-------+---------------+----+-------------------+
only showing top 20 rows
十、项目性能调优
1.集群优化
存储格式的选择:https://www.infoq.cn/article/bigdata-store-choose/
压缩格式的选择:
默认:snapy
.config("spark.sql.parquet.compression.codec","gzip")修改
2.代码优化
选择高性能算子
复用已有的数据
3.参数优化
并行度:
spark.sql.shuffle.partitions
200
配置在为联接或聚合进行数据洗牌时使用的分区数。
spark-submit:
--conf spark.sql.shuffle.partitions=500
IDEA:
.config("","")
分区字段类型推测:
spark.sql.sources.partitionColumnTypeInference.enabled
spark-submit:
--conf spark.sql.sources.partitionColumnTypeInference.enabled=false
IDEA:
.config("","")
262 |58.32.19.255 | |2017-05-11 00:38:01|
|http://www.imooc.com/video/4600 |video |4600 |465 |218.75.35.226 | |2017-05-11 17:38:16|
|http://www.imooc.com/video/4600 |video |4600 |833 |222.129.235.182| |2017-05-11 07:11:36|
|http://www.imooc.com/article/17895|article|17895|320 |222.129.235.182| |2017-05-11 19:25:04|
|http://www.imooc.com/article/17898|article|17898|460 |202.96.134.133 | |2017-05-11 15:14:28|
|http://www.imooc.com/article/17899|article|17899|389 |222.129.235.182| |2017-05-11 02:43:15|
±---------------------------------±------±----±------±--------------±—±------------------+
only showing top 20 rows
## 十、项目性能调优
### 1.集群优化
~~~markdown
存储格式的选择:https://www.infoq.cn/article/bigdata-store-choose/
压缩格式的选择:
默认:snapy
.config("spark.sql.parquet.compression.codec","gzip")修改
2.代码优化
选择高性能算子
复用已有的数据
3.参数优化
并行度:
spark.sql.shuffle.partitions
200
配置在为联接或聚合进行数据洗牌时使用的分区数。
spark-submit:
--conf spark.sql.shuffle.partitions=500
IDEA:
.config("","")
分区字段类型推测:
spark.sql.sources.partitionColumnTypeInference.enabled
spark-submit:
--conf spark.sql.sources.partitionColumnTypeInference.enabled=false
IDEA:
.config("","")