6 Replies Latest reply on May 15, 2018 3:03 AM by Intel Corporation

    java -version error when configuring TensorFlow




      I have been following this thread: How can I install the last version of tensorflow v1.8 or v1.7  to install TensorFlow, but I have run into the 'java version  not being detected' issue and attempting to export the path, as mentioned in the thread, is not solving the problem. Is there an easier way to install and work with TensorFlow?


      Here is the error I am getting:


      (iMaterialist) [u14063@c009 tensorflow]$ ./configure

      WARNING: Output base '/home/u14063/.cache/bazel/_bazel_u14063/24f6828cf341db0236432cb7b8ada71b' is on NFS. This may lead to surprising failures and undetermined behavior.

      Java version not detected while at least 1.8 is needed.

      Please set JAVA_HOME.


      And this is how my commands look after I use qsub -I and try to export the path:


      [u14063@c009-n023 ~]$ java -version

      openjdk version "1.8.0_131"

      OpenJDK Runtime Environment (build 1.8.0_131-b12)

      OpenJDK 64-Bit Server VM (build 25.131-b12, mixed mode)

      [u14063@c009-n023 ~]$ whereis java

      java: /usr/bin/java /usr/lib/java /etc/java /usr/share/java /usr/share/man/man1/java.1.gz

      [u14063@c009-n023 ~]$ export JAVA_HOME=/usr/bin/java


      These commands did not solve the issue and I am still getting the same errors when ./configure.


      Let me know if you need any other information.