View Full Version : Can connect to HDFS but cannot access the actual files (application freezes)

04-02-2013, 11:02 AM

I'm using Kettle 4.4.0 against a CDH3 cluster.

I can connect to the cluster alright and see folders and files, though when I select any file the application freezes (not responding) - no error message is visilble

Any ideas...?

Thank you!


04-24-2013, 09:00 AM

I resolved this and I believe this could be a BUG in the Hadoop File Input step:

- Connect to the cluster hdfs://myserver - Port 8020 - it connects OK, "Open from folder field" shows: hdfs://myserver:8020/
- Navigate to a route and select a file. "Open from folder" field shows: hdfs://myserver:8020/data/raw/logs/
- Click OK

The selected file path looks like: hdfs://myserver/data/raw/logs/mylog.log - if I click on "show file name" or "show file content" the application freezes (no error message).

The thing is... the port (8020) is gone: it should be hdfs://myserver:8020/data/raw/logs/mylog.log but the port it's not there: it's hdfs://myserver/data/raw/logs/mylog.log - if I add the port manually then it works.

Hope this helps