Issue Details (XML | Word | Printable)

Key: HADOOP-153
Type: New Feature New Feature
Status: Closed Closed
Resolution: Fixed
Priority: Major Major
Assignee: Sharad Agarwal
Reporter: Doug Cutting
Votes: 1
Watchers: 12
Operations

If you were logged in you would be able to see more operations.
Hadoop Common

skip records that fail Task

Created: 21/Apr/06 12:40 AM   Updated: 08/Jul/09 04:51 PM
Return to search
Component/s: None
Affects Version/s: 0.2.0
Fix Version/s: 0.19.0

Time Tracking:
Not Specified

File Attachments:
  File Name Date Attached Attached By Size
Text File Licensed for inclusion in ASF works 153_1.patch 2008-07-17 01:09 PM Sharad Agarwal 37 kB
Text File Licensed for inclusion in ASF works 153_2.patch 2008-07-22 11:47 AM Sharad Agarwal 57 kB
Text File Licensed for inclusion in ASF works 153_3.patch 2008-07-25 10:56 AM Sharad Agarwal 63 kB
Text File Licensed for inclusion in ASF works 153_4.patch 2008-07-25 11:59 AM Sharad Agarwal 63 kB
Text File Licensed for inclusion in ASF works 153_5.patch 2008-07-29 06:12 AM Sharad Agarwal 63 kB
Text File Licensed for inclusion in ASF works 153_6.patch 2008-08-04 01:30 PM Sharad Agarwal 68 kB
Text File Licensed for inclusion in ASF works 153_7.patch 2008-08-06 10:28 AM Sharad Agarwal 68 kB
Text File Licensed for inclusion in ASF works 153_8.patch 2008-08-11 06:41 AM Sharad Agarwal 68 kB
Text File Licensed for inclusion in ASF works 153_9.patch 2008-08-11 11:04 AM Sharad Agarwal 68 kB
Text File Licensed for inclusion in ASF works skipRecords_wip1.patch 2008-06-24 12:05 PM Enis Soztutar 30 kB
Issue Links:
Dependants
 
Reference
 

Hadoop Flags: Reviewed
Release Note: Introduced record skipping where tasks fail on certain records. (org.apache.hadoop.mapred.SkipBadRecords)
Resolution Date: 11/Aug/08 12:04 PM


 Description  « Hide
MapReduce should skip records that throw exceptions.

If the exception is thrown under RecordReader.next() then RecordReader implementations should automatically skip to the start of a subsequent record.

Exceptions in map and reduce implementations can simply be logged, unless they happen under RecordWriter.write(). Cancelling partial output could be hard. So such output errors will still result in task failure.

This behaviour should be optional, but enabled by default. A count of errors per task and job should be maintained and displayed in the web ui. Perhaps if some percentage of records (>50%?) result in exceptions then the task should fail. This would stop jobs early that are misconfigured or have buggy code.

Thoughts?

 All   Comments   Work Log   Change History   Subversion Commits      Sort Order: Ascending order - Click to sort in descending order
Doug Cutting made changes - 21/Apr/06 01:50 AM
Field Original Value New Value
Affects Version/s 0.2 [ 12310813 ]
Assignee Doug Cutting [ cutting ]
Doug Cutting made changes - 21/Apr/06 01:52 AM
Fix Version/s 0.2 [ 12310813 ]
Doug Cutting made changes - 04/May/06 04:12 AM
Fix Version/s 0.2 [ 12310813 ]
Fix Version/s 0.3 [ 12310930 ]
Doug Cutting made changes - 06/Jun/06 03:21 AM
Assignee Doug Cutting [ cutting ]
Fix Version/s 0.3.0 [ 12310930 ]
Doug Cutting made changes - 06/Jun/06 06:16 AM
Workflow jira [ 12361349 ] no reopen closed [ 12372922 ]
Doug Cutting made changes - 07/Jun/06 04:37 AM
Workflow no reopen closed [ 12372922 ] no-reopen-closed [ 12373254 ]
Doug Cutting made changes - 03/Aug/06 05:46 PM
Workflow no-reopen-closed [ 12373254 ] no-reopen-closed, patch-avail [ 12377453 ]
Doug Cutting made changes - 15/Dec/06 09:40 PM
Assignee Owen O'Malley [ owen.omalley ]
Devaraj Das made changes - 22/Feb/08 07:02 PM
Fix Version/s 0.17.0 [ 12312913 ]
Assignee Owen O'Malley [ owen.omalley ] Devaraj Das [ devaraj ]
Robert Chansler made changes - 25/Mar/08 03:03 AM
Fix Version/s 0.17.0 [ 12312913 ]
Sharad Agarwal made changes - 24/Jun/08 09:39 AM
Assignee Devaraj Das [ devaraj ] Sharad Agarwal [ sharadag ]
Enis Soztutar made changes - 24/Jun/08 12:05 PM
Attachment skipRecords_wip1.patch [ 12384579 ]
Sharad Agarwal made changes - 17/Jul/08 01:09 PM
Attachment 153_1.patch [ 12386292 ]
Sharad Agarwal made changes - 22/Jul/08 11:47 AM
Attachment 153_2.patch [ 12386616 ]
Sharad Agarwal made changes - 24/Jul/08 01:33 PM
Summary skip records that throw exceptions skip records that fail Task
Sharad Agarwal made changes - 25/Jul/08 10:56 AM
Attachment 153_3.patch [ 12386871 ]
Sharad Agarwal made changes - 25/Jul/08 11:07 AM
Link This issue blocks HADOOP-3828 [ HADOOP-3828 ]
Sharad Agarwal made changes - 25/Jul/08 11:16 AM
Link This issue blocks HADOOP-3829 [ HADOOP-3829 ]
Sharad Agarwal made changes - 25/Jul/08 11:28 AM
Fix Version/s 0.19.0 [ 12313211 ]
Status Open [ 1 ] Patch Available [ 10002 ]
Sharad Agarwal made changes - 25/Jul/08 11:59 AM
Attachment 153_4.patch [ 12386874 ]
Sharad Agarwal made changes - 29/Jul/08 06:12 AM
Attachment 153_5.patch [ 12387079 ]
Sharad Agarwal made changes - 29/Jul/08 06:13 AM
Status Patch Available [ 10002 ] Open [ 1 ]
Sharad Agarwal made changes - 29/Jul/08 06:14 AM
Status Open [ 1 ] Patch Available [ 10002 ]
Doug Cutting made changes - 31/Jul/08 08:15 PM
Status Patch Available [ 10002 ] Open [ 1 ]
Sharad Agarwal made changes - 04/Aug/08 01:30 PM
Attachment 153_6.patch [ 12387470 ]
Sharad Agarwal made changes - 05/Aug/08 06:01 AM
Status Open [ 1 ] Patch Available [ 10002 ]
Sharad Agarwal made changes - 06/Aug/08 10:28 AM
Attachment 153_7.patch [ 12387626 ]
Sharad Agarwal made changes - 11/Aug/08 06:41 AM
Attachment 153_8.patch [ 12387929 ]
Sharad Agarwal made changes - 11/Aug/08 11:04 AM
Attachment 153_9.patch [ 12387944 ]
Devaraj Das made changes - 11/Aug/08 12:04 PM
Resolution Fixed [ 1 ]
Hadoop Flags [Reviewed]
Status Patch Available [ 10002 ] Resolved [ 5 ]
Koji Noguchi made changes - 15/Aug/08 04:26 PM
Link This issue relates to HADOOP-3954 [ HADOOP-3954 ]
Sharad Agarwal made changes - 16/Oct/08 06:36 AM
Release Note If there are bugs in user code which lead to failure of Tasks on certain records, this feature can be used to skip these record on further attempts. Refer org.apache.hadoop.mapred.SkipBadRecords for various settings.
Robert Chansler made changes - 21/Oct/08 11:15 PM
Release Note If there are bugs in user code which lead to failure of Tasks on certain records, this feature can be used to skip these record on further attempts. Refer org.apache.hadoop.mapred.SkipBadRecords for various settings. Introduced record skipping for where tasks fail on certain records.
Robert Chansler made changes - 21/Oct/08 11:16 PM
Release Note Introduced record skipping for where tasks fail on certain records. Introduced record skipping where tasks fail on certain records. (org.apache.hadoop.mapred.SkipBadRecords)
Nigel Daley made changes - 20/Nov/08 11:38 PM
Status Resolved [ 5 ] Closed [ 6 ]
Owen O'Malley made changes - 08/Jul/09 04:51 PM
Component/s mapred [ 12310690 ]

Powered by a free Atlassian JIRA open source license for Apache Software Foundation. Try JIRA - bug tracking software for your team.
Atlassian JIRA the Professional Issue Tracker. (Enterprise Edition, Version: 3.13.2-#335) - Bug/feature request - Atlassian news - Contact Administrators