I am trying to use HDFS file system in sap BODS 4.2 and could not able to read the data from HDFS.
I gave my Namenode Host, port, root directory and file name correctly and still i could not able to view the data. (all my credentials are correct) Is there any additional things to be installed in my machine to get the data from HDFS…?? Kindly share your knowledge and correct me if am missing anything here…
Thanks in advance,
With Regards
Vijay kumar Vanamala.
I am trying to configure BODS and Hadoop. I have read the fantastic blog Connecting SAP DataServices to Hadoop Hive - SAP Community and all linked documents. Actually I am facing problems with the HDFS configuration. In designer I can create a HDFS file format and I can see the data inside it. When I put this file format in a dataflow I can also see the content of this file with the preview functionality BUT when I run the job I get an error message:
My Environment:
-BODS 4.2 on Linux Redhat
-Hadoop 2.6.0 client installed on the BODS server
The hadoop cluster where I am trying to connect to is a Hortonworks Hadoop cluster. The only difference between my installation procedure and the proposal in the blog is that I have installed the hadoop client on the BODS server manually and not via Ambari server.
I have spent a lot of time to trying to solve it but without success. Do you know what should I check? Do you have some more documentation about that? The Hive connector could be configured and it is working well.
I am facing the same issue. We have BODS installed on Linux in one of the Hadoop cluster. I can see the data from Designer which is installed on separate Windows Server. When i run the job i get Error in data flow " Failed to initialize HDFS. Check HDFS environment setup.
I didn’t see a resolution for this post, but I ran into the same error:
I was able to resolve this using the following:
A shared library is missing from the /usr/lib64 directory. Make a copy manually to that location from /usr/lib and create a symbolic link to the file in /usr/lib64.
To work around this issue, before re-running the job, enter the following on the job server machine where the master node of the Hadoop cluster is running:
cd /usr/lib
cp libhdfs.so.0.0.0 /usr/lib64
cd /usr/lib64
ln -s libhdfs.so.0.0.0 libhdfs.so
I then began getting this error:
<Error: HDFS Failed Connect to myservername>
I was able to resolve that issue by making sure I used the hostname in the Data Services Job and making sure my hostname was in my /etc/hosts file.
(BOB member since 2005-09-16)