-1

I did these steps: 

mkdir ~/glibc_install; cd ~/glibc_install
wget http://ftp.gnu.org/gnu/glibc/glibc-2.15.tar.gz
tar zxvf glibc-2.15.tar.gz
cd glibc-2.15
mkdir build
cd build
../configure --prefix=/opt/glibc-2.15
make -j4
sudo make install
export LD_LIBRARY_PATH=/opt/glibc-2.15/lib

Still ldd --version giving 2.12 as its version.

I am using tensorflow 1.15.0 java dependency in my application.while loading the model it is thorwing below given exception.

java.lang.UnsatisfiedLinkError: /tmp/tensorflow_native_libraries-1583474080231-0/libtensorflow_jni.so: /lib64/libc.so.6: version `GLIBC_2.15' not found

Nicktar
  • 5,548
  • 1
  • 28
  • 43
Senthil
  • 111
  • 1
  • 2
  • 11

1 Answers1

2

On CentOS, you can only upgrade glibc by upgrading to a newer CentOS version. Core system libraries are an integral part of the operating system, and if you replace them, you have a different operating system. At that point, you might as well just upgrade.

To solve your original problem, you need to use a build of Tensorflow for Red Hat Enterprise Linux 6 or CentOS 6.

For future reference, it would be interesting to know why you are using an operating system that was released nearly ten years ago for a compute-intense task (like most tasks involving Tensorflow are). If this is original hardware from 2010 or earlier, it is not very efficient. If the hardware has been upgraded since, why not the operating system as well? Typically, hardware replacement is an opportunity for migrating to a new operating system, too.

Florian Weimer
  • 32,022
  • 3
  • 48
  • 92