Skip to content
Discussion options

You must be logged in to vote

All five crashes fault at the same offset in libm.so.6, so this looks like a single glibc math routine (likely exp/log/pow, used by numpy/scikit-learn), not paperless-ngx code.

Your CPU (AMD Opteron 4386, Piledriver, no AVX2) has known CPUID quirks that can trick glibc's ifunc resolver into picking an AVX2 code path anyway, which would also explain why it's nondeterministic.

Try forcing glibc to skip that path:

GLIBC_TUNABLES=glibc.cpu.hwcaps=-AVX2,-AVX512F,-FMA

Set that as a container env var and restart. Should be a glibc/CPU-detection issue, not something fixable in paperless-ngx itself.

(Note: I used Claude to help investigate and draft this suggestion, worth confirming it actually r…

Replies: 5 comments

Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
0 replies
Answer selected by stumpylog
Comment options

You must be logged in to vote
0 replies
Comment options

You must be logged in to vote
0 replies
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
2 participants