[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441336 ] Arun C Murthy commented on HADOOP-489: -- I meant a) conceptually with b) being the actual implementation ... and yes, I was proposing exactly what you have in

[jira] Updated: (HADOOP-514) namenode heartbeat interval should be configurable

[ http://issues.apache.org/jira/browse/HADOOP-514?page=all ] Milind Bhandarkar updated HADOOP-514: - Status: Patch Available (was: Open) Fix Version/s: 0.8.0 Affects Version/s: 0.7.0 Attached patch. > namenode heartbeat interv

[jira] Updated: (HADOOP-514) namenode heartbeat interval should be configurable

[ http://issues.apache.org/jira/browse/HADOOP-514?page=all ] Milind Bhandarkar updated HADOOP-514: - Attachment: heartbeat.patch Mde heartbeat configurable, in seconds, though. > namenode heartbeat interval should be configurable > --

[jira] Updated: (HADOOP-594) Safemode default threshold should be 0.999

[ http://issues.apache.org/jira/browse/HADOOP-594?page=all ] Konstantin Shvachko updated HADOOP-594: --- Status: Patch Available (was: Open) I changed the default. > Safemode default threshold should be 0.999 > --

[jira] Updated: (HADOOP-594) Safemode default threshold should be 0.999

[ http://issues.apache.org/jira/browse/HADOOP-594?page=all ] Konstantin Shvachko updated HADOOP-594: --- Attachment: DefaultThreshold.patch > Safemode default threshold should be 0.999 > -- > > K

[jira] Assigned: (HADOOP-594) Safemode default threshold should be 0.999

[ http://issues.apache.org/jira/browse/HADOOP-594?page=all ] Konstantin Shvachko reassigned HADOOP-594: -- Assignee: Konstantin Shvachko > Safemode default threshold should be 0.999 > -- > >

[jira] Created: (HADOOP-594) Safemode default threshold should be 0.999

Safemode default threshold should be 0.999 -- Key: HADOOP-594 URL: http://issues.apache.org/jira/browse/HADOOP-594 Project: Hadoop Issue Type: Improvement Reporter: Konstantin Shvachko

[jira] Updated: (HADOOP-498) fsck should execute faster

[ http://issues.apache.org/jira/browse/HADOOP-498?page=all ] Owen O'Malley updated HADOOP-498: - Fix Version/s: 0.8.0 (was: 0.7.1) > fsck should execute faster > -- > > Key: HADOOP-498 >

[jira] Commented: (HADOOP-514) namenode heartbeat interval should be configurable

[ http://issues.apache.org/jira/browse/HADOOP-514?page=comments#action_12441304 ] Milind Bhandarkar commented on HADOOP-514: -- Work started by milind. > namenode heartbeat interval should be configurable > ---

[jira] Assigned: (HADOOP-514) namenode heartbeat interval should be configurable

[ http://issues.apache.org/jira/browse/HADOOP-514?page=all ] Milind Bhandarkar reassigned HADOOP-514: Assignee: Milind Bhandarkar > namenode heartbeat interval should be configurable > -- > >

[jira] Updated: (HADOOP-498) fsck should execute faster

[ http://issues.apache.org/jira/browse/HADOOP-498?page=all ] Milind Bhandarkar updated HADOOP-498: - Status: Patch Available (was: Open) Fix Version/s: 0.7.1 Affects Version/s: 0.7.0 Fixed. Fsck executes much faster now. > fsc

[jira] Updated: (HADOOP-498) fsck should execute faster

[ http://issues.apache.org/jira/browse/HADOOP-498?page=all ] Milind Bhandarkar updated HADOOP-498: - Attachment: fsck.patch Attached patch. fsck functionality is moved into namenode aand made available as a servlet. This avoids multiple roundtrips fr

[jira] Updated: (HADOOP-586) Job Name should not ben empty, if its is not given bu user, Hadoop should use original Jar name as the job name.

[ http://issues.apache.org/jira/browse/HADOOP-586?page=all ] Owen O'Malley updated HADOOP-586: - Fix Version/s: 0.8.0 (was: 0.7.1) > Job Name should not ben empty, if its is not given bu user, Hadoop should use > original Jar name

[jira] Updated: (HADOOP-592) NullPointerException in toString for RPC connection objects

[ http://issues.apache.org/jira/browse/HADOOP-592?page=all ] Owen O'Malley updated HADOOP-592: - Status: Patch Available (was: Open) > NullPointerException in toString for RPC connection objects > -

[jira] Updated: (HADOOP-593) NullPointerException in JobTracker's ExireTaskTracker thread

[ http://issues.apache.org/jira/browse/HADOOP-593?page=all ] Owen O'Malley updated HADOOP-593: - Status: Patch Available (was: Open) > NullPointerException in JobTracker's ExireTaskTracker thread >

[jira] Updated: (HADOOP-593) NullPointerException in JobTracker's ExireTaskTracker thread

[ http://issues.apache.org/jira/browse/HADOOP-593?page=all ] Owen O'Malley updated HADOOP-593: - Attachment: catch-exceptions-in-jt.patch This patch: 1. replaces the host name in the persistent job log with the task tracker name (for lost task tra

[jira] Created: (HADOOP-593) NullPointerException in JobTracker's ExireTaskTracker thread

NullPointerException in JobTracker's ExireTaskTracker thread Key: HADOOP-593 URL: http://issues.apache.org/jira/browse/HADOOP-593 Project: Hadoop Issue Type: Bug Componen

[jira] Updated: (HADOOP-592) NullPointerException in toString for RPC connection objects

[ http://issues.apache.org/jira/browse/HADOOP-592?page=all ] Owen O'Malley updated HADOOP-592: - Attachment: connection-null-ptr.patch This patch grabs the inet address when the connection is formed so that we don't get a null pointer exception when tryi

[jira] Created: (HADOOP-592) NullPointerException in toString for RPC connection objects

NullPointerException in toString for RPC connection objects --- Key: HADOOP-592 URL: http://issues.apache.org/jira/browse/HADOOP-592 Project: Hadoop Issue Type: Bug Affects Versions

[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441230 ] Owen O'Malley commented on HADOOP-489: -- I'm confused by your proposal in that a and b sound like alterantives. I think we should do: 1. write the user logs t

[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441225 ] Yoram Arnon commented on HADOOP-489: that's reasonable. so the cap is defined per job, with a default that can be overriden in the jobconf, and the cap per tas

[jira] Commented: (HADOOP-590) Reducer's pass merger should utilize temporary directories on different disks

[ http://issues.apache.org/jira/browse/HADOOP-590?page=comments#action_12441219 ] Runping Qi commented on HADOOP-590: --- I observed from my current running job that the throughput for the sortPass (sort map output files into runs) is a bit fast

[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441214 ] Sameer Paranjpye commented on HADOOP-489: - A global cap on the log size is feasible, but dividing by the number of nodes doesn't look right. Some nodes are

[jira] Created: (HADOOP-591) Reducer sort should even out the pass factors in merging different pass

Reducer sort should even out the pass factors in merging different pass --- Key: HADOOP-591 URL: http://issues.apache.org/jira/browse/HADOOP-591 Project: Hadoop Issue Type:

[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441208 ] Arun C Murthy commented on HADOOP-489: -- > seems that one log file per tasktracker, per job is better than one file per > task, no? Sorry, my bad, that's what

[jira] Commented: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=comments#action_12441197 ] Yoram Arnon commented on HADOOP-489: seems that one log file per tasktracker, per job is better than one file per task, no? logs of rogue tasks should be limit

[jira] Created: (HADOOP-590) Reducer's pass merger should utilize temporary directories on different disks

Reducer's pass merger should utilize temporary directories on different disks -- Key: HADOOP-590 URL: http://issues.apache.org/jira/browse/HADOOP-590 Project: Hadoop

[jira] Assigned: (HADOOP-489) Seperating user logs from system logs in map reduce

[ http://issues.apache.org/jira/browse/HADOOP-489?page=all ] Arun C Murthy reassigned HADOOP-489: Assignee: Arun C Murthy (was: Owen O'Malley) I'll take a stab at a reasonable set of requirements for a first-cut: a) One log file per task b) Define

[jira] Created: (HADOOP-589) Locality hints for Reduce

Locality hints for Reduce - Key: HADOOP-589 URL: http://issues.apache.org/jira/browse/HADOOP-589 Project: Hadoop Issue Type: New Feature Components: mapred Reporter: Benjamin Reed It would be n

[jira] Updated: (HADOOP-588) JobTracker History bug - kill() ed tasks are logged wrongly as finished.

[ http://issues.apache.org/jira/browse/HADOOP-588?page=all ] Sanjay Dahiya updated HADOOP-588: - Status: Patch Available (was: Open) > JobTracker History bug - kill() ed tasks are logged wrongly as finished. >

[jira] Updated: (HADOOP-588) JobTracker History bug - kill() ed tasks are logged wrongly as finished.

[ http://issues.apache.org/jira/browse/HADOOP-588?page=all ] Sanjay Dahiya updated HADOOP-588: - Attachment: Hadoop-588.patch This patch fixes 2 issues with jobtracker history logging. 1. JobInProgress.kill() now logs jobs as failed instead of finished.

[jira] Created: (HADOOP-588) JobTracker History bug - kill() ed tasks are logged wrongly as finished.

JobTracker History bug - kill() ed tasks are logged wrongly as finished. - Key: HADOOP-588 URL: http://issues.apache.org/jira/browse/HADOOP-588 Project: Hadoop Issue Ty