You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mahout.apache.org by paritosh ranjan <pa...@gmail.com> on 2012/10/10 23:00:18 UTC

seqdirectory and cluster-reuters.sh not working properly?

I tried running kmeans through cluster-reuters-sh on a pseudo distributed
hadoop cluster.
Kmeans fails, saying no clusters found.

After debugging a bit, I have found that the problem lies in seqdirectory.
The seqdirectory command is not creating the output folder
reuters-out-seqdir (neither on local file system, nor on hdfs).
Is anyone having idea of how to solve it? or is it a bug?

PS : I am using ubuntu 11.10

Re: seqdirectory and cluster-reuters.sh not working properly?

Posted by paritosh ranjan <pa...@gmail.com>.
There was one local and one mahout problem behind this.
I will just point them out in case it can help others.

1) I was not allocating enough memory to JVM, so seqdirectory was failing
(very silently) while running locally ( during the creation of
reuters-out-seqdir ).
2) After I fixed this, there was multiple slf4j binding problem. I have
fixed it, https://issues.apache.org/jira/browse/MAHOUT-1099 should take
care of this problem.

On Thu, Oct 11, 2012 at 2:31 AM, paritosh ranjan
<pa...@gmail.com>wrote:

> correction : seqdirectory is creating empty folders on hdfs.
>
>
> On Thu, Oct 11, 2012 at 2:30 AM, paritosh ranjan <
> paritoshranjan5@gmail.com> wrote:
>
>> I tried running kmeans through cluster-reuters-sh on a pseudo distributed
>> hadoop cluster.
>> Kmeans fails, saying no clusters found.
>>
>> After debugging a bit, I have found that the problem lies in
>> seqdirectory. The seqdirectory command is not creating the output folder
>> reuters-out-seqdir (neither on local file system, nor on hdfs).
>> Is anyone having idea of how to solve it? or is it a bug?
>>
>> PS : I am using ubuntu 11.10
>>
>>
>>
>

Re: seqdirectory and cluster-reuters.sh not working properly?

Posted by paritosh ranjan <pa...@gmail.com>.
correction : seqdirectory is creating empty folders on hdfs.

On Thu, Oct 11, 2012 at 2:30 AM, paritosh ranjan
<pa...@gmail.com>wrote:

> I tried running kmeans through cluster-reuters-sh on a pseudo distributed
> hadoop cluster.
> Kmeans fails, saying no clusters found.
>
> After debugging a bit, I have found that the problem lies in seqdirectory.
> The seqdirectory command is not creating the output folder
> reuters-out-seqdir (neither on local file system, nor on hdfs).
> Is anyone having idea of how to solve it? or is it a bug?
>
> PS : I am using ubuntu 11.10
>
>
>