fix: override NVIDIA entrypoint in autoresearch image#113
Merged
Conversation
The nvidia/cuda base images set ENTRYPOINT to nvidia_entrypoint.sh which enforces a strict CUDA driver version check. On RunPod machines with driver 550 (CUDA 12.4 max), this kills the container before /start.sh runs — pods enter an undebuggable boot loop with 0 uptime. Adding ENTRYPOINT ["/start.sh"] to the autoresearch Dockerfile only, so the fix is scoped and doesn't affect the shared base image or any downstream templates. Made-with: Cursor
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
ENTRYPOINT ["/start.sh"]to the autoresearch Dockerfile onlynvidia_entrypoint.shwhich kills the container on hosts with older NVIDIA drivers (550 = CUDA 12.4 max, but this image needs CUDA 12.8)Context
The autoresearch template (
x7o8gn1p4f) enters an undebuggable boot loop on RunPod machines with driver <570. The NVIDIA entrypoint fails the CUDA compatibility check and exits before/start.shruns. No logs, no SSH, just 0 uptime forever.This is the same approach ComfyUI uses — its image already has
ENTRYPOINT ["/start.sh"].Replaces #112 which tried to fix this in the base image but would break child images that override CMD.
Test plan
/workspace/autoresearchpopulatedtorch.cuda.is_available()returns TrueMade with Cursor