--- On Tue, 4/13/10, Dharani Selvaraj wrote:
From: Dharani Selvaraj
Subject: problem in assigning an array
To: k_gokulapr...@yahoo.com
Date: Tuesday, April 13, 2010, 4:16 PM
Hello,
While we are trying to assign the values to a 3-d array,in the
map function we have got
Todd Lipcon wrote:
Most likely a kernel bug. In previous versions of Debian there was a buggy
forcedeth driver, for example, that caused it to drop off the network in
high load. Who knows what new bug is in 2.6.32 which is brand spanking new.
Yes, it looks like it is a kernel bug alright (see t
I have a large text file, approximately 500mb containing key value pairs on
each line. I would like to implement Hadoop MapFile so that I can access any
key,value pair fairly quickly. To construct either the Reader or Writer the
MapFile requires a Configurations object and a File System object. I
Correction, they are 100Mbps NIC's...
iperf shows that we're getting about 95 Mbits/sec from one node to another.
On Apr 12, 2010, at 1:05 PM, Andrew Nguyen wrote:
> @Todd:
>
> I do need the sorting behavior, eventually. However, I'll try it with zero
> reduce jobs to see.
>
> @Alex:
>
> Ye
Hadoop Fans, we wanted to share some news with the Hadoop community
about new upcoming courses, new locations, and a substantial discount
on next week's session in the Bay Area.
We're excited to offer an extended sysadmin course and new courses on
Hive and HBase at this year's Hadoop Summit. You c
Andrew,
here are some tips for hadoop runtime config:
http://cloudepr.blogspot.com/2009/09/cluster-facilities-hardware-and.html
also
here are some results from my cluster (using 1GE NICs, Fiber), Dell 5500,
24GB, 8-core (16 hypervised), JBOD, i saw slightly better numbers on a
different 4-nodes c
also http://www.slideshare.net/cloudera/hw09-optimizing-hadoop-deployments
On Tue, Apr 13, 2010 at 12:58 PM, alex kamil wrote:
> Andrew,
>
> here are some tips for hadoop runtime config:
> http://cloudepr.blogspot.com/2009/09/cluster-facilities-hardware-and.html
> also
>
> here are some results
On Mon, Apr 12, 2010 at 1:45 PM, Andrew Nguyen <
andrew-lists-had...@ucsfcti.org> wrote:
> I don't think you can :-). Sorry, they are 100Mbps NIC's... I get
> 95Mbit/sec from one node to another with iperf.
>
> Should I still be expecting such dismal performance with just 100Mbps?
>
Yes - in my
On Tue, Apr 13, 2010 at 4:13 AM, stephen mulcahy
wrote:
> Todd Lipcon wrote:
>
>> Most likely a kernel bug. In previous versions of Debian there was a buggy
>> forcedeth driver, for example, that caused it to drop off the network in
>> high load. Who knows what new bug is in 2.6.32 which is brand
Good to know... The problem is that I'm in an academic environment that
needs a lot of convincing regarding new computational technologies. I need
to show proven benefit before getting the funds to actually implement
anything. These servers were the best I could come up with for this
proof-of-co
On Tue, Apr 13, 2010 at 11:40 AM, Andrew Nguyen <
andrew-lists-had...@ucsfcti.org> wrote:
> Good to know... The problem is that I'm in an academic environment that
> needs a lot of convincing regarding new computational technologies. I need
> to show proven benefit before getting the funds to ac
Hey Andrew,
I can name 3 California universities (San Diego, Caltech, Santa-Barbera) that
use Hadoop at a small (~20TB raw) or medium scale (~800TB raw). Why not go
talk to those guys?
Otherwise, you might just be able to confirm old hardware is old (there's good
money that you might be hard
I thought I saw a way to specify the block size for individual files using the
command-line using "hadoop dfs -put/copyFromLocal..." However, I can't seem to
find the reference anywhere.
I see that I can do it via the API but no references to a command-line
mechanism. Am I just remembering so
Hi,
Pass the -D property in command line. eg:
Hadoop fs -Ddfs.block.size= .
You can check if its actually set the way you needed by hadoop fs -stat %o
HTH,
Amogh
On 4/14/10 9:01 AM, "Andrew Nguyen" wrote:
I thought I saw a way to specify the block size for individual files using the
com
Hi,
The file system object will contain the scheme, authority etc for the given uri
or path. The conf object acts as reference ( unable to get a better terminology
) to this info.
Looking at the MapFileOutputFormat should help provide better understanding as
to how writers and readers are initia
Hi all,
I am using the Cloudera Hadoop distribution version 0.20.2+228.
I have a small 9 node cluster and when I try to stop the Hadoop DFS
and Mapred using
the stop-mapred.sh and stop-dfs.sh scripts, it downs shutdown some of
the TaskTrackers and DataNodes. I get a message saying no tasktracker
Hi Abhishek,
Are you using the tarball or the RPMs/debs? The issue is most likely that
your pid files are ending up in /tmp and thus getting cleaned out
periodically.
-Todd
On Tue, Apr 13, 2010 at 11:07 PM, abhishek sharma wrote:
> Hi all,
>
> I am using the Cloudera Hadoop distribution version
Hi Todd,
I am using the tarball.
Let me try configuring the pid files to stored somewhere else.
Thanks for the tip,
Abhishek
On Tue, Apr 13, 2010 at 11:10 PM, Todd Lipcon wrote:
> Hi Abhishek,
>
> Are you using the tarball or the RPMs/debs? The issue is most likely that
> your pid files are en
Hi Ted,
Were you referring to the Hadoop 0.20.2 distribution or the CDH version?
I just looked at the FairScheduler assignTasks function in Hadoop
dist. 0.20.2 and it is the same as version 0.20.0, and it will assign
only 1 Map and 1 reduce task to a tasktracker per heartbeat (as far I
can tell b
19 matches
Mail list logo