~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.
Python academiaanthropicarxivbravedeep-researchencryptionhome-automationhomeserverlocallocal-deep-researchlocal-llmmistralollamaopenaipubmedresearchresearch-toolretrieval-augmented-generationsearxngself-hosted
fix(docker): use exec form for HEALTHCHECK CMD (hadolint DL3025) (#5404)
LearningCircuit22b9de908f55
diff --git a/Dockerfile b/Dockerfileindex 46aa776665..8adb5c8d93 100644--- a/Dockerfile+++ b/Dockerfile@@ -372,14 +372,13 @@ RUN chmod +x /scripts/ollama_entrypoint.sh \EXPOSE 5000# Health check for container orchestration (Docker, Kubernetes, etc.)-# The ``timeout=8`` on urlopen is load-bearing: without it, urllib hangs forever-# on a slow/blocked server. Docker's --timeout=10s only SIGKILLs the ``sh -c``-# wrapper; the python child gets reparented to PID 1 and continues holding a-# TCP socket open against the app (one pidfd per hung child on PID 1). With-# ``timeout=8`` the child returns/raises before Docker's wall, exits cleanly,-# and gets reaped.+# Exec form (CMD [...]) runs python as the direct healthcheck PID, so+# Docker SIGKILLs it directly on --timeout with no ``sh -c`` wrapper to+# leak a reparented child onto PID 1. ``timeout=8`` stays below Docker's+# 10s wall so urlopen raises (python exits 1) cleanly instead of being+# SIGKILLed at 137.HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \- CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:5000/api/v1/health', timeout=8)" || exit 1+ CMD ["python", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:5000/api/v1/health', timeout=8)"]STOPSIGNAL SIGINT