hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kris Nuttycombe <kris.nuttyco...@gmail.com>
Subject Running into problems with DistributedCache
Date Thu, 15 Apr 2010 18:06:14 GMT
Hi, all,

I'm having problems with my Mapper instances accessing the
DistributedCache. A bit of background:

I'm running on a single-node cluster, just trying to get my first
map/reduce job functioning. Both the job tracker and the primary
namenode exist on the same host. In the client, I am able to
successfully add a file to the distributed cache, but when my Mapper
instance attempts to read the file it fails, despite the fact that the
path it fails on exists on the system where the job is running.

Here is a paste detailing the code where the error is occurring,
related log output from the node where the job runs, and filesystem
information from the same:

http://paste.pocoo.org/show/202242/

The failure appears to be originating from these lines in DFSClient.java

      LocatedBlocks newInfo = callGetBlockLocations(namenode, src, 0,
prefetchSize);
      if (newInfo == null) {
        throw new IOException("Cannot open filename " + src);
      }

I've attempted to trace back through the code to try to figure out why
newInfo might be null, but I quickly got lost. Can someone please help
me figure out why it can't find this file?

Thank you,

Kris

Mime
View raw message