default entrypoint

llama_stack/distribution/build_container.sh

@@ -102,7 +102,7 @@ add_to_docker <<EOF
 # This would be good in production but for debugging flexibility lets not add it right now
 # We need a more solid production ready entrypoint.sh anyway
 #
-CMD ["python", "-m", "llama_stack.distribution.server.server", "./llamastack-run.yaml"]
+ENTRYPOINT ["python", "-m", "llama_stack.distribution.server.server"]
 
 EOF
 

llama_stack/distribution/configure_container.sh

@@ -32,6 +32,9 @@ DOCKER_OPTS="$DOCKER_OPTS --security-opt label=disable"
 
 set -x
 $DOCKER_BINARY run $DOCKER_OPTS -it \
+  --entrypoint "/usr/local/bin/llama" \
   -v $host_build_dir:$container_build_dir \
   $docker_image \
-  llama stack configure ./llamastack-build.yaml --output-dir $container_build_dir
+  "stack" \
+  "configure" \
+  "./llamastack-build.yaml"

llama_stack/distribution/server/server.py

@@ -408,7 +408,11 @@ async def resolve_impls_with_routing(run_config: StackRunConfig) -> Dict[Api, An
     return impls, specs
 
 
-def main(yaml_config: str, port: int = 5000, disable_ipv6: bool = False):
+def main(
+    yaml_config: str = "./llamastack-run.yaml",
+    port: int = 5000,
+    disable_ipv6: bool = False,
+):
     with open(yaml_config, "r") as fp:
         config = StackRunConfig(**yaml.safe_load(fp))