Fix sky start for AMI without conda (skypilot-org#1532)

* fix no conda img

* fix tpu yaml

* add yaml

* rename

* rename

* Update examples/tests/test_yamls/no_conda_ami.yaml

Co-authored-by: Zhanghao Wu <[email protected]>

Co-authored-by: Zhanghao Wu <[email protected]>

Author: infwinston

examples/tests/test_yamls/no_conda_ami.yaml

@@ -0,0 +1,9 @@
+resources:
+  cloud: aws
+  instance_type: g4dn.xlarge
+  region: us-west-2
+  image_id: ami-0fe5af21074ad2a10  # Deep learning AMI with CUDA 11.6 without conda installed
+
+run: |
+  nvidia-smi
+

examples/tpu/tpuvm_mnist.yaml

@@ -20,6 +20,7 @@ setup: |
     pip install "jax[tpu]>=0.2.16" -f https://storage.googleapis.com/jax-releases/libtpu_releases.html
     pip install --upgrade clu
     pip install -e flax
+    pip install tensorflow tensorflow-datasets
   fi
 
 

sky/templates/aws-ray.yml.j2

@@ -120,7 +120,7 @@ setup_commands:
     pip3 --version > /dev/null 2>&1 || (curl -sSL https://bootstrap.pypa.io/get-pip.py -o get-pip.py && python3 get-pip.py && echo "PATH=$HOME/.local/bin:$PATH" >> ~/.bashrc);
     (type -a python | grep -q python3) || echo 'alias python=python3' >> ~/.bashrc;
     (type -a pip | grep -q pip3) || echo 'alias pip=pip3' >> ~/.bashrc;
-    (which conda > /dev/null 2>&1 && conda init > /dev/null && conda config --set auto_activate_base false) || (wget -nc https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh && bash Miniconda3-latest-Linux-x86_64.sh -b && eval "$(~/miniconda3/bin/conda shell.bash hook)" && conda init && conda config --set auto_activate_base true);
+    (which conda > /dev/null 2>&1 && conda init > /dev/null) || (wget -nc https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh && bash Miniconda3-latest-Linux-x86_64.sh -b && eval "$(~/miniconda3/bin/conda shell.bash hook)" && conda init && conda config --set auto_activate_base true);
     source ~/.bashrc;
     (pip3 list | grep ray | grep {{ray_version}} 2>&1 > /dev/null || pip3 install -U ray[default]=={{ray_version}}) && mkdir -p ~/sky_workdir && mkdir -p ~/.sky/sky_app;
     (pip3 list | grep skypilot && [ "$(cat {{sky_remote_path}}/current_sky_wheel_hash)" == "{{sky_wheel_hash}}" ]) || (pip3 uninstall skypilot -y; pip3 install "$(echo {{sky_remote_path}}/{{sky_wheel_hash}}/skypilot-{{sky_version}}*.whl)[aws]" && echo "{{sky_wheel_hash}}" > {{sky_remote_path}}/current_sky_wheel_hash || exit 1);

tests/test_smoke.py

@@ -271,6 +271,24 @@ def test_image_id_dict_with_zone():
     run_one_test(test)
 
 
+def test_image_no_conda():
+    name = _get_cluster_name()
+    test = Test(
+        'image_no_conda',
+        [
+            # Use image id dict.
+            f'sky launch -y -c {name} examples/tests/test_yamls/no_conda_ami.yaml',
+            f'sky logs {name} 1 --status',
+            f'sky stop {name} -y',
+            f'sky start {name} -y',
+            f'sky exec {name} examples/tests/test_yamls/no_conda_ami.yaml',
+            f'sky logs {name} 2 --status',
+        ],
+        f'sky down -y {name}',
+    )
+    run_one_test(test)
+
+
 def test_stale_job():
     name = _get_cluster_name()
     test = Test(