org.apache.hadoop.hive.ql.exec.errors

Class RegexErrorHeuristic

java.lang.Object

org.apache.hadoop.hive.ql.exec.errors.RegexErrorHeuristic

All Implemented Interfaces:

ErrorHeuristic

Direct Known Subclasses:

DataCorruptErrorHeuristic, MapAggrMemErrorHeuristic, ScriptErrorHeuristic

public abstract class RegexErrorHeuristic
extends java.lang.Object
implements ErrorHeuristic

Simple heuristic where the query and the lines of the task log file are run through regular expressions to see if they resemble a known error condition. Only a single regular expression can be supplied to match the query whereas multiple regular expressions can be supplied to match lines from the log file. A mapping is maintained from the regular expression to the lines from the log file that it matched.

Constructor Summary

Constructors

Constructor and Description
RegexErrorHeuristic()

Method Summary

Modifier and Type	Method and Description
protected JobConf	getConf()
abstract ErrorAndSolution	getErrorAndSolution() Examine the hive query, job configuration, and the lines from the task log seen so far though processLogLine() and generate a possible cause/solution.
protected java.util.Set<java.lang.String>	getLogRegexes()
protected boolean	getQueryMatches()
protected java.lang.String	getQueryRegex()
protected java.util.Map<java.lang.String,java.util.List<java.lang.String>>	getRegexToLogLines()
void	init(java.lang.String query, JobConf conf) Initialize this error heuristic.
void	processLogLine(java.lang.String line) Process the given log line.
protected void	reset() Resets to state before any processLogLine() calls.
protected void	setQueryRegex(java.lang.String queryRegex)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RegexErrorHeuristic

public RegexErrorHeuristic()

Method Detail

setQueryRegex

protected void setQueryRegex(java.lang.String queryRegex)

getQueryRegex

protected java.lang.String getQueryRegex()

getQueryMatches

protected boolean getQueryMatches()

getLogRegexes

protected java.util.Set<java.lang.String> getLogRegexes()

getRegexToLogLines

protected java.util.Map<java.lang.String,java.util.List<java.lang.String>> getRegexToLogLines()

getConf

protected JobConf getConf()

init

public void init(java.lang.String query,
                 JobConf conf)

Description copied from interface: ErrorHeuristic

Initialize this error heuristic. Must be called before any other methods are called

Specified by:

init in interface ErrorHeuristic

getErrorAndSolution

public abstract ErrorAndSolution getErrorAndSolution()

Description copied from interface: ErrorHeuristic

Examine the hive query, job configuration, and the lines from the task log seen so far though processLogLine() and generate a possible cause/solution. Once this method is called, the implementing class should be reset to the state before any processLogLine() calls were made.

Specified by:

getErrorAndSolution in interface ErrorHeuristic

Returns:

a matching error, or null if a suitable match wasn't found.

processLogLine

public void processLogLine(java.lang.String line)

Description copied from interface: ErrorHeuristic

Process the given log line. It should be called for every line in the task log file, in sequence.

Specified by:

processLogLine in interface ErrorHeuristic

reset

protected void reset()

Resets to state before any processLogLine() calls.