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

RuntimeError: Error in dlopen #250

Closed
abastola0 opened this issue Apr 30, 2024 · 1 comment
Closed

RuntimeError: Error in dlopen #250

abastola0 opened this issue Apr 30, 2024 · 1 comment

Comments

@abastola0
Copy link

abastola0 commented Apr 30, 2024

I faced this issue while running pytorch application but some similar issues were mentioned for julia as well (referenced here: #JuliaLang/julia#49246)

RuntimeError: Error in dlopen: libtorch_cuda_linalg.so: cannot open shared object file: No such file or directory

Here is the debug output:

     36333:	file=libtorch_cuda_linalg.so [0];  dynamically loaded by /lib/libdlfaker.so [0]
     36333:	find library=libtorch_cuda_linalg.so [0]; searching
     36333:	 search path=/home/abastol/modules/anaconda3/envs/newsn/bin/../lib		(RPATH from file python)
     36333:	  trying file=/home/abastol/modules/anaconda3/envs/newsn/bin/../lib/libtorch_cuda_linalg.so
     36333:	 search cache=/etc/ld.so.cache
     36333:	 search path=/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu:/lib:/usr/lib		(system search path)
     36333:	  trying file=/lib/x86_64-linux-gnu/libtorch_cuda_linalg.so
     36333:	  trying file=/usr/lib/x86_64-linux-gnu/libtorch_cuda_linalg.so
     36333:	  trying file=/lib/libtorch_cuda_linalg.so
     36333:	  trying file=/usr/lib/libtorch_cuda_linalg.so
     36333:	
Traceback (most recent call last):
  File "/home/abastol/working/currentAnomalyDetection/src/test.py", line 16, in <module>

The first hit was correct (trying file=home/abastol/modules/anaconda3/envs/newsn/bin/../lib/libtorch_cuda_linalg.so) but the not sure about the second one. There were two windows opened so not sure if the second hit (trying file=/usr/lib/libtorch_cuda_linalg.so)was for that (some concurrency issue perhaps).

Anyway a manual export to the conda env library path seems to fix this issue (referenced here:#JuliaLang/julia#49246) but this might be temporary.

@dcommander
Copy link
Member

I believe this is a duplicate of #107. If so, it is a known issue, and the workaround is documented in the VirtualGL User's Guide. (The workaround is just a convenience option to vglrun that does exactly what you describe: adding the application's library directory to LD_LIBRARY_PATH. Refer to #107 and the VirtualGL User's Guide for an explanation of why that is necessary.)

@dcommander dcommander closed this as not planned Won't fix, can't repro, duplicate, stale May 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants