Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Problems and solutions in installing tf on Redhat #2924

Closed
jinfengfeng opened this issue Jun 17, 2016 · 6 comments
Closed

Problems and solutions in installing tf on Redhat #2924

jinfengfeng opened this issue Jun 17, 2016 · 6 comments

Comments

@jinfengfeng
Copy link

jinfengfeng commented Jun 17, 2016

I encountered several problems in installing tensorflow on Redhat system. I solved them by googling solutions. Hope my experience may be helpful to some users.

OS version: Red Hat Enterprise Linux Server release 6.2 (Santiago)
Linux kernel: 2.6.32-220.el6.x86_64

Problems and solutions:

1. glibc version is too low

Install higer version glibc, e.g. glibc-2.17.
Please use at least version 2.17. You can configure the installation using --prefix option to specify the path to install, and then modify LD_LIBRARY_PATH to include the new glibc library.

2. link problem after install higher version of glibc

Error info looks like:
"
error while loading shared libraries: __vdso_time: invalid mode for dlopen():Invalid argument
"
You need to use the following method to run a program (eg. python):
/path/to/glibc-2.17/lib/ld-linux-x86-64.so.2 --library-path /path/to/glibc-2.17/lib:$LD_LIBRARY_PATH:/path/to/gcc-5.2.0/lib64:/usr/lib64/ /path/to/anaconda/bin/python2.7

3. numpy version problem

Install new version of numpy using pip.

4. protobuf3 not found, and
Error in python after 'import tensorflow': TypeError: init() got an unexpected keyword argument 'syntax'

pip install 'protobuf>=3.0.0a3'

5. failed call to cuInit: CUDA_ERROR_NO_DEVICE
Explicitly specify a cuda device in you environment variable:
export CUDA_VISIBLE_DEVICES=0

6. libcuda.so.1 not found

Explicitly specify the path of this lib in your LD_LIBRARY_PATH environment variable.
libcuda.so is in usually in /usr/lib64

7. ELFCLASS32 error
elf is incorrect, should use a 64-bits .so file (usually in */lib64)

8. glibc 2.15 cannot create regular file `/var/db/Makefile': Permission denied
Use glibc 2.17 or even higher version.

@sherrym
Copy link
Contributor

sherrym commented Jun 17, 2016

Thanks, @jinfengfeng .

@sherrym sherrym closed this as completed Jun 17, 2016
@PapaMadeleine2022
Copy link

@jinfengfeng I can not understand the second point 2. link problem after install higher version of glibc in your above solution. How to solve this error:
while loading shared libraries: __vdso_time: invalid mode for dlopen():Invalid argument
?

@naveenjafer
Copy link

@jinfengfeng I did not understand Point 2 either, could you please help. @IvyGongoogle Were you able to solve the issue?

@shm007g
Copy link

shm007g commented Dec 12, 2017

I did not understand Point 2 either, could you please help. @IvyGongoogle Were you able to solve the issue?

@naveenjafer
Copy link

naveenjafer commented Dec 12, 2017

I tried what was mentioned in point 2, but running a command like
/opt/glibc-2.17/lib/ld-linux-x86-64.so.2 --library-path /opt/glibc-2.17/lib:$LD_LIBRARY_PATH:usr/lib64/ /home/njafer/python/bin/python3 at least works, but a command like
/opt/glibc-2.17/lib/ld-linux-x86-64.so.2 --library-path /opt/glibc-2.17/lib:$LD_LIBRARY_PATH:usr/lib64/ ls gives me a
ls: error while loading shared libraries: ls: cannot open shared object file

@mjwen
Copy link
Contributor

mjwen commented Dec 22, 2017

@naveenjafer
To run ls, you may need something like

/opt/glibc-2.17/lib/ld-linux-x86-64.so.2 --library-path /opt/glibc-2.17/lib:$LD_LIBRARY_PATH:usr/lib64/ /bin/ls

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants