Re: [jira] Commented: (HADOOP-286) copyFromLocal throws LeaseExpiredException

2006-07-20 Thread Konstantin Shvachko
The problem with increasing the lease period is that in case of a task failure the task will retry, and start creating the file it needs to create from scratch, which won't be possible since the file under this name is still locked. So the task will need to wait 5 minutes instead of 1 to start t

[jira] Commented: (HADOOP-281) dfs.FSDirectory.mkdirs can create sub-directories of a file!

2006-07-20 Thread Konstantin Shvachko (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-281?page=comments#action_12422504 ] Konstantin Shvachko commented on HADOOP-281: We can still return the correct boolean value to preserve public interface, but throw an exception at the s

[jira] Updated: (HADOOP-286) copyFromLocal throws LeaseExpiredException

2006-07-20 Thread Konstantin Shvachko (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-286?page=all ] Konstantin Shvachko updated HADOOP-286: --- Attachment: Lease.patch This is a very simple patch that renews leases only when pendingCreates is not empty. This prevents the client from sendi

[jira] Commented: (HADOOP-260) the start up scripts should take a command line parameter --config making it easy to run multiple hadoop installation on same machines

2006-07-20 Thread Milind Bhandarkar (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-260?page=comments#action_12422487 ] Milind Bhandarkar commented on HADOOP-260: -- I have started looking at this issue. The original patch contains duplicate shell-script in all scripts. We co

[jira] Commented: (HADOOP-281) dfs.FSDirectory.mkdirs can create sub-directories of a file!

2006-07-20 Thread Yoram Arnon (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-281?page=comments#action_12422485 ] Yoram Arnon commented on HADOOP-281: throwing an exception was suggested, but rejected in order not to change a public interface that may be (is) used by appli

[jira] Commented: (HADOOP-344) TaskTracker passes incorrect file path to DF under cygwin

2006-07-20 Thread Konstantin Shvachko (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-344?page=comments#action_12422474 ] Konstantin Shvachko commented on HADOOP-344: This patch is getting stale. Is there any reason for not committing this? I'm getting FD related exceptions

[jira] Commented: (HADOOP-372) should allow to specify different inputformat classes for different input dirs for Map/Reduce jobs

2006-07-20 Thread Runping Qi (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-372?page=comments#action_12422473 ] Runping Qi commented on HADOOP-372: --- Second thought about the stuff related to InputFormat after chatting with Owen. The getSplits should stay as non-static, bu

[jira] Commented: (HADOOP-54) SequenceFile should compress blocks, not individual entries

2006-07-20 Thread Arun C Murthy (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-54?page=comments#action_12422470 ] Arun C Murthy commented on HADOOP-54: - Sounds good... looks like we're all on the same page; I'll get going on this. Appreciate both of you spending time... I'l

[jira] Commented: (HADOOP-372) should allow to specify different inputformat classes for different input dirs for Map/Reduce jobs

2006-07-20 Thread Runping Qi (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-372?page=comments#action_12422467 ] Runping Qi commented on HADOOP-372: --- Doug, My thought is to add a Map object to the JobConf class that keep track the explicit association between input path a

[jira] Created: (HADOOP-374) native support for gzipped text files

2006-07-20 Thread Yoram Arnon (JIRA)
native support for gzipped text files -- Key: HADOOP-374 URL: http://issues.apache.org/jira/browse/HADOOP-374 Project: Hadoop Issue Type: New Feature Components: mapred Reporter: Yor

[jira] Commented: (HADOOP-281) dfs.FSDirectory.mkdirs can create sub-directories of a file!

2006-07-20 Thread eric baldeschwieler (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-281?page=comments#action_12422458 ] eric baldeschwieler commented on HADOOP-281: Why not throw an exception on failure? That seems like the more correct thing to do. Checking return co

Re: [jira] Commented: (HADOOP-286) copyFromLocal throws LeaseExpiredException

2006-07-20 Thread Eric Baldeschwieler
Why not significantly extend the lease period as well, to say 5 minutes and have well behaved clients release the lease explicitly as soon as they can? Clients could then try to renew starting at say 2.5 minutes and try every 30 seconds til 4.5 minutes have expired... Seems like this woul

[jira] Commented: (HADOOP-54) SequenceFile should compress blocks, not individual entries

2006-07-20 Thread eric baldeschwieler (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-54?page=comments#action_12422454 ] eric baldeschwieler commented on HADOOP-54: --- I meant zero compressed. So we're all on the same page. (Of course recording key & value lengths rather than

[jira] Commented: (HADOOP-54) SequenceFile should compress blocks, not individual entries

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-54?page=comments#action_12422393 ] Doug Cutting commented on HADOOP-54: Just to be clear, delta compression means something different to me than zero-compression. The former represents a lists o

[jira] Resolved: (HADOOP-237) Standard set of Performance Metrics for Hadoop

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-237?page=all ] Doug Cutting resolved HADOOP-237. - Fix Version/s: 0.5.0 Resolution: Fixed I just committed this. It would be great to get some, e.g., Ganglia screenshots up on the wiki demoing this st

[jira] Commented: (HADOOP-371) ant tar should package contrib jars

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-371?page=comments#action_12422386 ] Doug Cutting commented on HADOOP-371: - > I thought that the way you get all the code on the CLASSPATH in bin/hadoop: > is by placing the contrib jar-s along wit

[jira] Commented: (HADOOP-372) should allow to specify different inputformat classes for different input dirs for Map/Reduce jobs

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-372?page=comments#action_12422385 ] Doug Cutting commented on HADOOP-372: - Can you provide more details? Is the intent for the mapred.input.format.class property to become multivalued, a paralle

[jira] Updated: (HADOOP-54) SequenceFile should compress blocks, not individual entries

2006-07-20 Thread Arun C Murthy (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-54?page=all ] Arun C Murthy updated HADOOP-54: Attachment: VIntCompressionResults.txt Here's a test run analysing compression of VInts (exported from recordio). Essentially we get almost 50% savings (either

[jira] Commented: (HADOOP-370) TaskTracker startup fails if any mapred.local.dir entries don't exist

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-370?page=comments#action_12422378 ] Doug Cutting commented on HADOOP-370: - Yes, let's cache the "good dirs". If a drive goes offline or becomes unwritable while a node is running, then we should

[jira] Commented: (HADOOP-54) SequenceFile should compress blocks, not individual entries

2006-07-20 Thread Doug Cutting (JIRA)
[ http://issues.apache.org/jira/browse/HADOOP-54?page=comments#action_12422374 ] Doug Cutting commented on HADOOP-54: Re documentation: perhaps we should add file-format documentation to the javadoc as a part of this change. This could look