本文试图实现的需求场景为：以学习Spark知识点为目的，编写Scala利用Spark解析800M的tomcat日志文件，打印一段时间内ERROR级别记录的前10行，统计每分钟的日志记录数，并将统计结果存入mysql数据库中。之前曾用JAVA写过一次同样的处理逻辑，但在学习了Scala之后，真的感觉在计算方面Scala要比JAVA方便的多。没有学习Scala语言的同学速度速度了啊……
技术要点
- 将日志文件写入HDFS中，相对路径PATH为“nova.log”
- 注意JAVA堆栈异常日志的处理
- 将解析后的异常日志全部存到SparkSQL中或Hive数据仓库中
- 通过编写SQL查询一段时间内ERROR级别记录的前10行
- 统计每分钟的日志记录数，并将统计结果存入mysql数据库中,便于上层应用直接使用计算结果
解析前后对比
解析前：
解析后：
解析代码
LoggerApp.scala：
import java.text.SimpleDateFormat
import java.util.Date
import org.apache.spark.SparkConf
import org.apache.spark.SparkContext
import org.apache.spark.rdd.RDD
import org.apache.spark.rdd.RDD.rddToPairRDDFunctions
import org.apache.spark.sql.Row
import org.apache.spark.sql.SQLContext
import org.apache.spark.sql.hive.HiveContext
import org.apache.spark.sql.types.StringType
import org.apache.spark.sql.types.StructField
import org.apache.spark.sql.types.StructType
import org.apache.spark.sql.types.TimestampType
/**
* 日志解析
*/
object LoggerApp {
def main(args: Array[String]): Unit = {
println("<!--开始解析-->")
val reg = "^(\\d{4}-\\d{2}-\\d{2} \\d{2}:\\d{2}:\\d{2}.\\d{3}) (\\[.*\\]) (.*) (.*) - ([\\s\\S]*)$"
val path = "nova.log"
val sc = new SparkContext(new SparkConf().setAppName("日志解析"))
val textRDD = sc.textFile(path)
/**
* 处理一条日志包括多行的情况
*/
var key = ""
val formatRDD = textRDD.map { x =>
if (x.matches(reg)) {
key = x
Pair.apply(key, "")
} else {
Pair.apply(key, x)
}
}.reduceByKey((a, b) => { a + "\n" + b }).map(x => x._1 + x._2)
/**
* 将字符串转换为Logger
*/
val loggerRDD: RDD[Logger] = formatRDD.map { x =>
{
val reg.r(time, thread, level, logger, msg) = x //通过正则取值
val log = new Logger(formatDate(time), thread, level, logger, msg)
log
}
}.cache()
/**
* TODO 通过类的反射机制来定义数据库Scheme，但在scala语言中不知道为啥就是不成功，此处浪费了许久留着以后研究吧
*/
/*val sqlc = new SQLContext(sc)
sqlc.createDataFrame(loggerRDD, classOf[Logger]).registerTempTable("logger")*/
/**
* 定义数据库Scheme
*/
val schemaString = "time thread level logger msg"
val schema =
StructType(
schemaString.split(" ").map(fieldName =>
if ("time".equals(fieldName))
StructField(fieldName, TimestampType, true)
else
StructField(fieldName, StringType, true)))
/**
* 将Logger转换为Row
*/
val rowRDD = loggerRDD.map { log =>
Row(
formatDate(log.time),
log.thread,
log.level,
log.logger,
log.msg)
}
/**
* 利用SQL进行查询过滤
*/
// val sqlc = bySQLContext(sc, rowRDD, schema);
val sqlc = byHiveContext(sc, rowRDD, schema);
val df = sqlc.sql("select * from logger where level='ERROR' and time between '2016-03-21 11:00:00' and '2016-03-21 12:00:00' order by time")
val errLogRDD = df.map { x =>
new Logger(
formatDate(x.getTimestamp(0)),
x.getString(1),
x.getString(2),
x.getString(3),
x.getString(4))
}
for (log <- errLogRDD.take(10)) {
println("time:" + formatDateToStr(log.time))
println("thread:" + log.thread)
println("level:" + log.level)
println("logger:" + log.logger)
println("msg:" + log.msg)
}
println("<!--解析结束-->")
}
/**
* 创建临时表
*/
def bySQLContext(sc: SparkContext, rowRDD: RDD[Row], schema: StructType): SQLContext = {
val sqlc = new SQLContext(sc)
sqlc.createDataFrame(rowRDD, schema).registerTempTable("logger")
sqlc
}
/**
* 创建永久表，需要提前搭建好Spark与Hive的集成环境
*/
def byHiveContext(sc: SparkContext, rowRDD: RDD[Row], schema: StructType): SQLContext = {
val sqlc = new HiveContext(sc)
sqlc.sql("drop table if exists logger")
sqlc.sql("CREATE TABLE IF NOT EXISTS logger (time TIMESTAMP, thread STRING, level STRING, logger STRING, msg STRING)")
sqlc.createDataFrame(rowRDD, schema).write.mode("overwrite").saveAsTable("logger")
sqlc
}
def formatDate(str: String): Date = {
new SimpleDateFormat("yyyy-MM-dd HH:mm:ss.SSS").parse(str)
}
def formatDate(timestamp: java.sql.Timestamp): Date = {
new SimpleDateFormat("yyyy-MM-dd HH:mm:ss.SSS").parse(timestamp.toString())
}
def formatDate(date: Date): java.sql.Timestamp = {
new java.sql.Timestamp(date.getTime)
}
def formatDateToStr(date: Date): String = {
new SimpleDateFormat("yyyy-MM-dd HH:mm:ss.SSS").format(date)
}
}
Logger.scala：
import java.util.Date
class Logger extends Serializable {
var time: Date = null
var thread: String = ""
var level: String = ""
var logger: String = ""
var msg: String = ""
def this(time: Date, thread: String, level: String, logger: String, msg: String) {
this()
this.time = time;
this.thread = thread;
this.level = level;
this.logger = logger;
this.msg = msg;
}
}
统计并写入Mysql
LoggerMysqlApp.scala：
import org.apache.spark.SparkContext
import org.apache.spark.SparkConf
import org.apache.spark.sql.hive.HiveContext
import org.apache.spark.sql.SQLContext
import java.util.Date
import java.text.SimpleDateFormat
import org.apache.spark.sql.types.StringType
import org.apache.spark.sql.types.StructType
import org.apache.spark.sql.types.StructField
import org.apache.spark.sql.types.IntegerType
import org.apache.spark.sql.Row
import java.util.Properties
object LoggerMysqlApp {
def main(args: Array[String]): Unit = {
val sc = new SparkContext(new SparkConf().setAppName("输出写入Mysql"))
/**
* 从hive中加载数据
*/
val hivec = new HiveContext(sc)
val df = hivec.sql("select * from logger")
val loggerRDD = df.rdd.map { x =>
new Logger(
LoggerApp.formatDate(x.getTimestamp(0)),
x.getString(1),
x.getString(2),
x.getString(3),
x.getString(4))
}
val resultRDD = loggerRDD.map { logger =>
Pair(formatDateToStr(logger.time), 1)
}.reduceByKey((a, b) =>
{ a + b }).map(f =>
Row(f._1, f._2)).sortBy(f => f.getInt(1), false, 2)
for (r <- resultRDD.take(10)) {
println(r.getString(0) + ":" + r.getInt(1))
}
/**
* 定义数据库Scheme
*/
val schemaString = "time count"
val schema =
StructType(
schemaString.split(" ").map(fieldName =>
if ("time".equals(fieldName))
StructField(fieldName, StringType, true)
else
StructField(fieldName, IntegerType, true)))
/**
* TODO计算每分钟日志的个数
*/
val connectionProperties = new Properties()
connectionProperties.setProperty("user", "root")
connectionProperties.setProperty("password", ".")
new SQLContext(sc).createDataFrame(resultRDD, schema).write.jdbc(
"jdbc:mysql://192.168.136.128:3306/logger",
"logger",
connectionProperties);
}
def formatDateToStr(date: Date): String = {
new SimpleDateFormat("yyyy-MM-dd HH:mm").format(date)
}
}
