Revert "Merge gcsfue into main (#317)" (#369)

This reverts commit c4e64e4.

Author: pawloch00

.github/workflows/build_tests.yaml

@@ -35,15 +35,13 @@ env:
   TPU_CLUSTER_NAME: build-xpk-2-nodepools
   WORKLOAD_NAME: xpktest-build-${{ github.run_attempt }}
   PATHWAYS_WORKLOAD_NAME: xpkpw-build-${{ github.run_attempt }}
-  RUN_ID: "pr-${{ github.event.number }}"
+  CLUSTER_ARGUMENTS: "--network=${{secrets.NETWORK_NAME}} --subnetwork=${{secrets.SUBNETWORK_NAME}} --maintenance-window=23:50"
   PROJECT_ID: ${{secrets.PROJECT_NAME}}
   A3_MEGA_TEST_CLUSTER_NAME: "xpk-mega-ctk-int"
   A3_ULTRA_TEST_CLUSTER_NAME: "xpk-ultra-ctk-int"
   GKE_ML_TEST_CLUSTER_NAME: "xpk-gke-ml"
   ZONE: us-central2-a
   REGION: us-central2
-  STORAGE_NAME: test-storage
-  PW_CLUSTER_ARGUMENTS: "--network=${{secrets.NETWORK_NAME}} --subnetwork=${{secrets.SUBNETWORK_NAME}} --maintenance-window=23:50"
 
 jobs:
   run-unit-tests:
@@ -134,10 +132,6 @@ jobs:
       with:
         version: '>= 363.0.0'
         install_components: 'beta,gke-gcloud-auth-plugin'
-    - name: Generate random seed
-      run: |
-        RANDOM_SEED=$((RANDOM % 10000))  # Generate a random number between 0 and 9999
-        echo "RANDOM_SEED=$RANDOM_SEED" >> $GITHUB_ENV
     - name: Verify gcp setup
       run: gcloud info
     - name: Set Google Cloud CLI properties to a unused zone to verify --zone arg is passed properly in commands.
@@ -152,56 +146,31 @@ jobs:
       run: gcloud components install kubectl && gcloud components install gke-gcloud-auth-plugin
     - name: Check xpk installation
       run: xpk --help
-    - name: Create a private Pathways-enabled XPK Cluster with 2x v4-8 nodepools. Larger num-nodes to avoid master resizing.
-      run: |
-        python3 xpk.py cluster create-pathways --cluster $TPU_CLUSTER_NAME --private --tpu-type=v4-8  --num-slices=2 \
-        --zone=us-central2-b --default-pool-cpu-machine-type=n1-standard-16 --default-pool-cpu-num-nodes=16 \
-        --reservation='${{ secrets.GCP_TPU_V4_RESERVATION }}' --enable-workload-identity --enable-gcsfuse-csi-driver --custom-cluster-arguments="${PW_CLUSTER_ARGUMENTS}"\
+    - name: Create a private Pathways-enabled XPK Cluster with 2x $TPU_TYPE nodepools. Larger num-nodes to avoid master resizing.
+      run: python xpk.py cluster create-pathways --cluster $TPU_CLUSTER_NAME --private --tpu-type=$TPU_TYPE  --num-slices=2 --zone=us-central2-b --default-pool-cpu-machine-type=n1-standard-16 --default-pool-cpu-num-nodes=16 --reservation='${{ secrets.GCP_TPU_V4_RESERVATION }}' --custom-cluster-arguments="${CLUSTER_ARGUMENTS}"
+    - name: Verify the created cluster is private
+      run: gcloud container clusters describe $TPU_CLUSTER_NAME --region=us-central2 --format="value(privateClusterConfig.enablePrivateNodes)" | grep 'True' || (echo 'The created cluster is not private.' && exit 1)
     - name: List out the nodepools on the cluster
       run: python xpk.py cluster describe --cluster $TPU_CLUSTER_NAME --zone=us-central2-b | grep -P '^(?=.*NODEPOOL_NAME)(?=.*SLICE)(?=.*TYPE)(?=.*EXPECTED_HEALTHY_NODES)(?=.*ACTUAL_HEALTHY_NODES)(?=.*TOTAL_NODES)'
     - name: Authenticate Docker
       run: gcloud auth configure-docker --quiet
-    - name: Create auto-mount Storage instance
-      run: |
-        python3 xpk.py storage create $STORAGE_NAME --cluster=$TPU_CLUSTER_NAME --zone=us-central2-b --type=gcsfuse \
-        --auto-mount=true \
-        --mount-point='/test-mount-point' --readonly=false --manifest='./tests/data/pv-pvc-templates.yaml'
-    - name: List and verify existing Storages
-      run: python3 xpk.py storage list --cluster $TPU_CLUSTER_NAME --zone=us-central2-b | tee output.txt | grep 'test-storage' || (echo 'No storage found' && cat output.txt && exit 1)
     - name: Create test script to execute in workloads
-      run: |
-        echo -e \
-        '#!/bin/bash \n
-        echo "Hello world from a test script!"
-        cd ~/../test-mount-point && echo "Hello world from a Github Action CI/CD test script!" > '$RANDOM_SEED'.txt' \
-        > test.sh
-    - name: Create a test script to execute in jobs.
-      run: echo -e '#!/bin/bash \nsleep 3 \necho "Hello world from a test script!"' > job_test.sh
+      run: echo -e '#!/bin/bash \nsleep 3 \necho "Hello world from a test script!"' > workload.sh
     - name: Run a job
       run: |
         python xpk.py run --cluster $TPU_CLUSTER_NAME --zone=us-central2-b workload.sh | awk '/Starting log streaming for pod xpk-def-app-profile-slurm-[a-zA-Z0-9]+-[0-9]+-[a-zA-Z0-9]+.../,/Job logs streaming finished./ { print }'
-    # TODO: fix Delete the job failing step
-    # - name: Delete the job
-    #   run: |
-    #     JOB_NAME=$(python3 xpk.py job ls --cluster $TPU_CLUSTER_NAME --zone=us-central2-b | grep 'xpk-def-app-profile-slurm-' | head -1 | awk '{print $1}')
-    #     kubectl delete job ${JOB_NAME}
-    - name: Run a base-docker-image workload
+    - name: Delete the job
       run: |
-        python3 xpk.py workload create --cluster $TPU_CLUSTER_NAME --workload $WORKLOAD_NAME  --command "bash test.sh"  \
-        --tpu-type=v4-8 --num-slices=2 --zone=us-central2-b
+        JOB_NAME=$(python3 xpk.py job ls --cluster $TPU_CLUSTER_NAME --zone=us-central2-b | grep 'xpk-def-app-profile-slurm-' | head -1 | awk '{print $1}')
+        kubectl delete job ${JOB_NAME}
+    - name: Run a base-docker-image workload
+      run: python xpk.py workload create --cluster $TPU_CLUSTER_NAME --workload $WORKLOAD_NAME  --command "bash workload.sh"  --tpu-type=$TPU_TYPE --num-slices=2 --zone=us-central2-b
     - name: Run xpk inspector with the workload created above
       run: python3 xpk.py inspector --cluster $TPU_CLUSTER_NAME --zone=us-central2-b  --workload $WORKLOAD_NAME
     - name: Wait for workload completion and confirm it succeeded
       run: python3 xpk.py workload list --cluster $TPU_CLUSTER_NAME --zone=us-central2-b --wait-for-job-completion $WORKLOAD_NAME --timeout 300
     - name: Run a Pathways workload on Ubuntu base image
-      run: |
-        python3 xpk.py workload create-pathways --cluster $TPU_CLUSTER_NAME --workload $PATHWAYS_WORKLOAD_NAME \
-        --docker-image='marketplace.gcr.io/google/ubuntu2004' --tpu-type=v4-8 --num-slices=2 --zone=us-central2-b \
-        --command "echo \"Hello world from a test script! \""
-    - name: Verify if the file was created in the GCS bucket
-      run: gsutil cp gs://xpk-ci-cd-tests/$RANDOM_SEED.txt .
-    - name: Check if the file contains desired content
-      run: grep 'Hello world from a Github Action CI/CD test script!' $RANDOM_SEED.txt
+      run: python xpk.py workload create-pathways --cluster $TPU_CLUSTER_NAME --workload $PATHWAYS_WORKLOAD_NAME --docker-image='marketplace.gcr.io/google/ubuntu2004' --tpu-type=$TPU_TYPE --num-slices=2 --zone=us-central2-b --command "echo \"Hello world from a test script! \""
     - name: Wait for Pathways workload completion and confirm it succeeded
       run: python3 xpk.py workload list --cluster $TPU_CLUSTER_NAME --zone=us-central2-b --wait-for-job-completion $PATHWAYS_WORKLOAD_NAME --timeout 300
     - name: List out the workloads on the cluster
@@ -212,10 +181,6 @@ jobs:
       run: python3 xpk.py workload delete --workload $WORKLOAD_NAME --cluster $TPU_CLUSTER_NAME --zone=us-central2-b
     - name: Delete the Pathways workload on the cluster
       run: python3 xpk.py workload delete --workload $PATHWAYS_WORKLOAD_NAME --cluster $TPU_CLUSTER_NAME --zone=us-central2-b
-    - name: Delete created GCS file
-      run: gsutil rm gs://xpk-ci-cd-tests/$RANDOM_SEED.txt
-    - name: Delete existing Storage
-      run: python3 xpk.py storage delete $STORAGE_NAME  --cluster $TPU_CLUSTER_NAME --zone=us-central2-b
     - name: Create test script to execute in batch
       run: echo -e '#!/bin/bash \n#SBATCH --unknown-flag=value\n echo "Hello world from a test script!"' > batch.sh
     - name: Run a batch job on the cluster

.github/workflows/lint_and_format.yml

@@ -22,7 +22,7 @@ on:
 
 jobs:
   build-and-test:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     strategy:
       matrix:
         python-version: ["3.10", "3.11"]

README.md

@@ -56,10 +56,7 @@ and the following GPU types:
 and the following CPU types:
 * n2-standard-32
 
-xpk also supports Google Cloud Storage solutions:
-* [Cloud Storage FUSE](https://cloud.google.com/storage/docs/gcs-fuse)
-
-# Permissions needed on Cloud Console:
+# Cloud Console Permissions on the user or service account needed to run XPK:
 
 * Artifact Registry Writer
 * Compute Admin
@@ -164,8 +161,6 @@ cleanup with a `Cluster Delete`.
 If you have failures with workloads not running, use `xpk inspector` to investigate
 more.
 
-If you need your Workloads to have persistent storage, use `xpk storage` to find out more.
-
 ## Cluster Create
 
 First set the project and zone through gcloud config or xpk arguments.
@@ -440,58 +435,14 @@ Currently, the below flags/arguments are supported for A3-Mega and A3-Ultra mach
   * --on-demand (only A3-Mega)
 
 
-## Storage
-Currently xpk supports Cloud Storage FUSE. A FUSE adapter that lets you mount and access Cloud Storage buckets as local file systems, so applications can read and write objects in your bucket using standard file system semantics.
-
-To use the GCS FUSE with XPK user needs to:
-- create a [Storage Bucket](https://pantheon.corp.google.com/storage/)
-- create a manifest with PersistentVolume and PersistentVolumeClaim that mounts to the Bucket. To learn how to properly
-set up PersistentVolume and PersistentVolumeClaim visit [GKE Cloud Storage documentation](https://cloud.google.com/kubernetes-engine/docs/how-to/persistent-volumes/cloud-storage-fuse-csi-driver#provision-static)
-
-Once it's ready user can define:
-
-`--type` - defines a type of a storage, currently xpk supports `gcsfuse` only.
-`--auto-mount` - if set to true means that all workloads should have a given storage mounted by default.
-`--mount-point` - defines the path on which a given storage should be mounted for a workload.
-`--manifest` - defines the path to manifest which contains PersistentVolume and PersistentVolumeClaim definitions
-`--readonly` - if set to true, workload can only read from storage.
-
-* Create a simple Storage
-
-    ```shell
-    python3 xpk.py storage create test-storage --project=$PROJECT
-    --cluster=xpk-test --type=gcsfuse --auto-mount=false \
-    --mount-point='/test-mount-point' --readonly=false \
-    --manifest='pv-pvc-auto-mount.yaml'
-    ```
-
-* Create a simple Workload with Storage attached
-    ```shell
-    python3 xpk.py workload create \
-    --workload xpk-test-workload --command "echo goodbye" \
-    --cluster xpk-test \
-    --tpu-type=v5litepod-16 \
-    --storage test-storage --projet=$PROJECT
-    ```
-
-* List Storage
-    ```shell
-    python3 xpk.py storage list --cluster xpk-test --zone=us-central2-b --projet=$PROJECT
-    ```
-
-* Delete Storage
-    ```shell
-    python3 xpk.py storage delete test-storage  --cluster xpk-test --zone=us-central2-b --projet=$PROJECT
-    ```
-
 ## Workload Create
 *   Workload Create (submit training job):
 
     ```shell
     python3 xpk.py workload create \
     --workload xpk-test-workload --command "echo goodbye" \
     --cluster xpk-test \
-    --tpu-type=v5litepod-16 --projet=$PROJECT
+    --tpu-type=v5litepod-16
     ```
 
 *   Workload Create for Pathways:

docs/contributing.md

@@ -45,7 +45,6 @@ source $VENV_DIR/bin/activate
 
 ```shell
 git clone https://github.com/google/xpk.git
-cd xpk
 pip install .[dev]
 ```
 

pyproject.toml

@@ -35,9 +35,6 @@ dependencies = [
   "ruamel.yaml",
   "pyyaml",
   "docker",
-  "kubernetes==31.0.0",
-  "google-cloud",
-  "google-api-core",
   "packaging"
 ]
 
@@ -67,9 +64,8 @@ dev = [
 version = {attr = "xpk.core.core.__version__"}
 
 [tool.setuptools]
-packages = ["xpk", "xpk.parser", "xpk.core", "xpk.commands", "xpk.api", "xpk.templates", "xpk.utils", "xpk.core.blueprint", "xpk.core.workload_decorators"]
+packages = ["xpk", "xpk.parser", "xpk.core", "xpk.commands", "xpk.utils", "xpk.core.blueprint", "xpk.core.workload_decorators"]
 package-dir = {"" = "src"}
-package-data = {"xpk.api" = ["storage_crd.yaml"], "xpk.templates" = ["storage.yaml"]}
 
 [tool.pyink]
 # Formatting configuration to follow Google style-guide.

pytype-conf.cfg

@@ -20,7 +20,6 @@ exclude =
     src/xpk/core/kjob.py
     src/xpk/core/kueue.py
     src/xpk/core/nap.py
-    src/xpk/core/storage.py
     src/xpk/core/pathways.py
     src/xpk/core/system_characteristics.py
     src/xpk/parser
@@ -29,7 +28,7 @@ exclude =
 # Keep going past errors to analyze as many files as possible.
 keep_going = True
 
-jobs = 5
+jobs = 10
 
 # Platform (e.g., "linux", "win32") that the target code runs on.
 platform = linux