This is the mapred-site.xml. I am using hadoop-0.20.1. The new queue, cqueue
is being displayed in jobtracker UI. But the job just does not get submitted
to it. It still goes to default queue.
mapred.job.tracker
cluster1:9001
mapred.jobtracker.taskScheduler
org.apache.hadoop.mapred.Capacit
Yes i did. But it still is the same. I tried removing the deafult queue
entirely. But that gives an error.
On Wed, Dec 16, 2009 at 1:47 PM, chaitanya krishna <
chaitanyavv.ii...@gmail.com> wrote:
> Hi,
>
> After updating conf/mapred-site.xml, did you restart using
> bin/stop-mapred.sh and bin/st
Hi,
for 20.1 version , the property name is mapred.job.queue.name while for 21,
it is mapreduce.job.queue.name.
Hope this helps.
- Chaitanya.
On Wed, Dec 16, 2009 at 1:54 PM, anjali nair wrote:
> Yes i did. But it still is the same. I tried removing the deafult queue
> entirely. But that gives
Thanks very much.
We use Ubuntu 8.0.4, and the kernel is 2.6.24-16-generic.
When we run jps, it shows:
12405 startup.jar
5950 startup.jar
19216 SecondaryNameNode
25381 Jps
19053 NameNode
19289 JobTracker
When I pass "-o private" for fuse-dfs just as: fuse_dfs
dfs://10.15.62.4:54310 /mnt/dfs -opriv
2009/12/15 William Kinney
> I had a similar issue and found the profiling information to be helpful:
> http://hadoop.apache.org/common/docs/current/mapred_tutorial.html#Profiling
>
> It should tell you where it's spending most of the time.
>
>
Thank you William, which problem did you have if you
Hey Weiming,
We've recently found a race condition in FUSE-DFS that can be triggered when
you run it from a host where GNOME is also running. It causes a segfault like
you report below.
The patch attached works for 19.1. Another option is to shut down all the
instances of GNOME on your compu
Hi,
First, I would like to apologise if this question has been asked before
(I am quite sure it has been) and I would appreciate very much if
someone replies with a link to the answer.
My question is quite simple.
I have to files or datasets having a list of integers.
example:
dataset A: (a
Hi, Brian, that is the point, you are right. thanks very much.
We used the Ubuntu Desktop for easy programing.
It appears that I have to upgrade the hadoop to the latest version.
On Wed, Dec 16, 2009 at 8:39 PM, Brian Bockelman wrote:
> Hey Weiming,
>
> We've recently found a race condition in
Sigh, I suppose this means I need to come out from under the rock I've been
hiding under and file a JIRA.
As I mentioned, the work-around is to mount FUSE-DFS when GNOME is not running.
Brian
On Dec 16, 2009, at 8:00 AM, Weiming Lu wrote:
> Hi, Brian, that is the point, you are right. thanks
Hi,
we just encountered some problems when restarting our namenode. I'd really
appreciate if anyone has any clue of what is going on here.
The error message is as follows:
09/12/16 14:25:03 INFO namenode.NameNode: STARTUP_MSG:
/
STARTUP_
Hi Eguzki,
Is one of the tables vastly smaller than the other? If one is small enough
to fit in RAM, you can do this like so:
1. Add the small file to the DistributedCache
2. In the configure() method of the mapper, read the entire file into an
ArrayList or somesuch in RAM
3. Set the input path o
Hi Fu-Ming,
Looks similar to this bug:
http://issues.apache.org/jira/browse/HDFS-686
Does this problem persist, or was it a one time occurrence?
-Todd
On Tue, Dec 15, 2009 at 5:42 PM, Fu-Ming Tsai wrote:
> Hello, all,
>
> I tried to execute 2 SecondaryNamenode in my env. However, one worked
Hi Erik,
A few things to try:
- does this FS store sensitive data or would it be possible to bzip2 the
files and upload them somewhere?
- can you add logging to the replay of FSEditLog so as to be aware of what
byte offset is causing the issue?
- DO take a backup of all of the state, immediately,
Thanks Todd,
That was my plan-B or workaround. Anyway, I am happy to see there is no
straight way to do so I could miss.
The "small" list is a list of userId (dim table), so I can assume it as
"small" but that can be a limitation in the scalability of our system. I
will test the upper limits
Hi Eguzki,
I wouldn't say the size of the list fitting into RAM would be the
scalability bottleneck. If you're doing a full cartesian join of your users
against a larger table, the fact that you're doing the full cartesian join
is going to be the bottleneck first :)
-Todd
On Wed, Dec 16, 2009 at
On Wed, Dec 16, 2009 at 12:29 PM, Todd Lipcon wrote:
> Hi Eguzki,
>
> I wouldn't say the size of the list fitting into RAM would be the
> scalability bottleneck. If you're doing a full cartesian join of your users
> against a larger table, the fact that you're doing the full cartesian join
> is go
Thanks Amogh. I solve it now. The reason is exactly what you said, I didn't
consume all the records in the reducer. I break the loop when meet certain
record. In this case, the rest records which I ignore will not be counted. So,
there is no problem at all!
-Gang
- 原始邮件
发件人: Amogh
Thanks a lot for your reply, Todd. I added some answers below.
On Wed, Dec 16, 2009 at 6:06 PM, Todd Lipcon wrote:
> Hi Erik,
>
> A few things to try:
>
> - does this FS store sensitive data or would it be possible to bzip2 the
> files and upload them somewhere?
>
Unfortunately, I think it'd be
On Wed, Dec 16, 2009 at 1:03 PM, Erik Bernhardsson wrote:
> Thanks a lot for your reply, Todd. I added some answers below.
>
> On Wed, Dec 16, 2009 at 6:06 PM, Todd Lipcon wrote:
>
> > Hi Erik,
> >
> > A few things to try:
> >
> > - does this FS store sensitive data or would it be possible to bzi
Hi All,
I am interested to know that can we use hadoop for applications where they
need more control over the data and it can specify which node will do which
part of the processing or the storage. For instance, suppose that I have two
data files (datasets, say 1 and 2) and setup a hadoop with two
Our machines suffered from bad memcpy performance, which became
apparent after profiling (a lot of time in System.arraycopy() for our
inputformat reader).
On 12/16/09, Dmitriy Lyfar wrote:
> 2009/12/15 William Kinney
>
>> I had a similar issue and found the profiling information to be helpful:
>
it sounds to me like you might want to split what you want to do up
into two separate jobs entirely... i don't quite understand your use
case since the point of hadoop is to spread your load as much (and
haphazardly!) as possible.
-mike
On Wed, Dec 16, 2009 at 3:43 PM, Ahmad Ali Iqbal
wrote:
>
Hi all, I'm wondering whether Amazon starts to support the newest stable
version of Hadoop, or we can still just use 0.18.3?
Song Liu
Last time I checked EMR only runs 0.18.3. You can use EC2 though, which
winds up being cheaper anyways.
On Wed, Dec 16, 2009 at 8:51 PM, 松柳 wrote:
> Hi all, I'm wondering whether Amazon starts to support the newest stable
> version of Hadoop, or we can still just use 0.18.3?
>
> Song Liu
>
You can build your own clusters on EC2, using Cloudera's distribution. It
worked for me.
Mark
On Wed, Dec 16, 2009 at 8:17 PM, Ed Kohlwey wrote:
> Last time I checked EMR only runs 0.18.3. You can use EC2 though, which
> winds up being cheaper anyways.
>
> On Wed, Dec 16, 2009 at 8:51 PM, 松柳 w
Hi Mike,
My understanding is, in hadoop job scheduling is done implicitly as you said
it spread load as much as possible. However, I want to control task
assignments to nodes. Let me put in a context of ad-hoc networking
application scenario where a mobile devices broadcast *Hello* packets
periodi
Tanx a lot finally..its working
--
Anjali M
Thanks for your help, Todd,
It happens whenever I restart SecondaryNamenode.
But what I'm curious is why the other host can work fine. Did you
encounter this problem?
Best regards,
Fu-Ming
On Thu, Dec 17, 2009 at 12:59 AM, Todd Lipcon wrote:
> Hi Fu-Ming,
>
> Looks similar to this bug:
>
> h
So, Can we use one machine which is not the namenode or datanode to
mount FUSE-DFS which is start up at console instead of X Window ?.
Yesterday, I installed the hadoop-0.18.2 on another machine, and
builded the FUSE-DFS successfully.
Should the conf files such as hadoop-site.xml, masters and sla
29 matches
Mail list logo