|
|
-
Reading Sequence File from Hadoop Distributed Cache ..
Mark Olimpiati 2012-10-10, 23:45
Hi,
I'm storing sequence files in the distributed cache which seems to be stored somewher under each node's /tmp .../local/archive/ ... path.
In mapper code, I tried using SequenceFile.Reader with all possible configurations (locally, distribtued) however, it can't find it. Are sequence files supported in the distribued cache??
Thank you, Mark
-
Re: Reading Sequence File from Hadoop Distributed Cache ..
Harsh J 2012-10-11, 02:33
Hi Mark,
DistributedCache files, when accessed from a Task, exist on the local file system. You should make sure the SequenceFile.Reader tries to read it with a LocalFS than a HDFS instance.
On Thu, Oct 11, 2012 at 5:15 AM, Mark Olimpiati <[EMAIL PROTECTED]> wrote: > Hi, > > I'm storing sequence files in the distributed cache which seems to be > stored somewher under each node's /tmp .../local/archive/ ... path. > > In mapper code, I tried using SequenceFile.Reader with all possible > configurations (locally, distribtued) however, it can't find it. Are > sequence files supported in the distribued cache?? > > Thank you, > Mark
-- Harsh J
-
Re: Reading Sequence File from Hadoop Distributed Cache ..
Mark Olimpiati 2012-10-11, 19:53
Thanks for the reply Harsh, but as I said I tried locally too by using the following:
FileSystem localFs = cachedFiles[0].getFileSystem(new Configuration()); Isn't the above supposed to give me the local file system ?? If yes, I tried that but didn't work, if not, can you show one example?
Thank you, Mark
On Wed, Oct 10, 2012 at 7:33 PM, Harsh J <[EMAIL PROTECTED]> wrote:
> Hi Mark, > > DistributedCache files, when accessed from a Task, exist on the local > file system. You should make sure the SequenceFile.Reader tries to > read it with a LocalFS than a HDFS instance. > > On Thu, Oct 11, 2012 at 5:15 AM, Mark Olimpiati <[EMAIL PROTECTED]> > wrote: > > Hi, > > > > I'm storing sequence files in the distributed cache which seems to be > > stored somewher under each node's /tmp .../local/archive/ ... path. > > > > In mapper code, I tried using SequenceFile.Reader with all possible > > configurations (locally, distribtued) however, it can't find it. Are > > sequence files supported in the distribued cache?? > > > > Thank you, > > Mark > > > > -- > Harsh J >
-
Re: Reading Sequence File from Hadoop Distributed Cache ..
Steve Loughran 2012-10-12, 18:27
On 11 October 2012 20:53, Mark Olimpiati <[EMAIL PROTECTED]> wrote:
> Thanks for the reply Harsh, but as I said I tried locally too by using the > following: > > FileSystem localFs = cachedFiles[0].getFileSystem(new > Configuration()); > > > Isn't the above supposed to give me the local file system ?? If yes, I > tried that but didn't work, if not, can you show one example? > > Thank you, > Mark gives you the default FS for that process, which can be different.
what about FileSystem.get("file://tmp",new Configuration())
|
|