Ignore zombie processes when detecting TorchServe status (#166)

* Ignore processes that are not running when detecting TorchServe status

* Detect zombie processes and ignore them before calling cmdline API

* Update PT DLC framework version to 2.1.0 and 2.2.0

* Update instance type to ensure newer CUDA driver version

* upgrade DLAMI version for tests

* Revert "Update instance type to ensure newer CUDA driver version"

This reverts commit bba00bd.

Author: namannandan

buildspec.yml

@@ -2,7 +2,7 @@ version: 0.2
 
 env:
   variables:
-    FRAMEWORK_VERSIONS: '2.0.0 2.0.1'
+    FRAMEWORK_VERSIONS: '2.1.0 2.2.0'
     CPU_INSTANCE_TYPE: 'ml.c4.xlarge'
     GPU_INSTANCE_TYPE: 'ml.g4dn.12xlarge'
     ECR_REPO: 'sagemaker-test'
@@ -49,12 +49,12 @@ phases:
       - $(aws ecr get-login --registry-ids $DLC_ACCOUNT --no-include-email --region $AWS_DEFAULT_REGION)
       - create-key-pair
 
-      # launch remote GPU instance with Deep Learning AMI GPU PyTorch 1.9 (Ubuntu 20.04)
+      # launch remote GPU instance with Deep Learning AMI GPU PyTorch 2.2 (Ubuntu 20.04)
       # build DLC GPU image because the base DLC image is too big and takes too long to build as part of the test
       - |
         for FRAMEWORK_VERSION in $FRAMEWORK_VERSIONS;
           do
-            launch-ec2-instance --instance-type $instance_type --ami-name ami-03e3ef8c92fdb39ad;
+            launch-ec2-instance --instance-type $instance_type --ami-name ami-081c4092fbff425f0;
             DLC_GPU_TAG="$FRAMEWORK_VERSION-dlc-gpu-$BUILD_ID";
             build_dir="test/container/$FRAMEWORK_VERSION";
             docker build -f "$build_dir/Dockerfile.dlc.gpu" -t $PREPROD_IMAGE:$DLC_GPU_TAG --build-arg region=$AWS_DEFAULT_REGION .;

src/sagemaker_pytorch_serving_container/torchserve.py

@@ -184,6 +184,9 @@ def _retrieve_ts_server_process():
     ts_server_processes = list()
 
     for process in psutil.process_iter():
+        if process.status() == psutil.STATUS_ZOMBIE:
+            continue
+
         if TS_NAMESPACE in process.cmdline():
             ts_server_processes.append(process)
 

test/conftest.py

@@ -52,7 +52,7 @@ def pytest_addoption(parser):
     parser.addoption('--instance-type')
     parser.addoption('--docker-base-name', default='sagemaker-pytorch-inference')
     parser.addoption('--region', default='us-west-2')
-    parser.addoption('--framework-version', default="2.0.0")
+    parser.addoption('--framework-version', default="2.1.0")
     parser.addoption('--py-version', choices=['2', '3'], default='3')
     parser.addoption('--processor', choices=['gpu', 'cpu'], default='cpu')
     # If not specified, will default to {framework-version}-{processor}-py{py-version}

test/container/2.0.0/Dockerfile.dlc.cpu renamed to test/container/2.1.0/Dockerfile.dlc.cpu

@@ -1,5 +1,5 @@
 ARG region
-FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.0.0-cpu-py310-ubuntu20.04-sagemaker
+FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.1.0-cpu-py310-ubuntu20.04-sagemaker
 
 COPY dist/sagemaker_pytorch_inference-*.tar.gz /sagemaker_pytorch_inference.tar.gz
 

test/container/2.0.0/Dockerfile.dlc.gpu renamed to test/container/2.1.0/Dockerfile.dlc.gpu

@@ -1,5 +1,5 @@
 ARG region
-FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.0.0-gpu-py310-cu118-ubuntu20.04-sagemaker
+FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.1.0-gpu-py310-cu118-ubuntu20.04-sagemaker
 
 COPY dist/sagemaker_pytorch_inference-*.tar.gz /sagemaker_pytorch_inference.tar.gz
 

test/container/2.0.1/Dockerfile.dlc.cpu renamed to test/container/2.2.0/Dockerfile.dlc.cpu

@@ -1,5 +1,5 @@
 ARG region
-FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.0.1-cpu-py310-ubuntu20.04-sagemaker
+FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.2.0-cpu-py310-ubuntu20.04-sagemaker
 
 COPY dist/sagemaker_pytorch_inference-*.tar.gz /sagemaker_pytorch_inference.tar.gz
 

test/container/2.0.1/Dockerfile.dlc.gpu renamed to test/container/2.2.0/Dockerfile.dlc.gpu

@@ -1,5 +1,5 @@
 ARG region
-FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.0.1-gpu-py310-cu118-ubuntu20.04-sagemaker
+FROM 763104351884.dkr.ecr.$region.amazonaws.com/pytorch-inference:2.2.0-gpu-py310-cu118-ubuntu20.04-sagemaker
 
 COPY dist/sagemaker_pytorch_inference-*.tar.gz /sagemaker_pytorch_inference.tar.gz