Revert "Modified parameters for BigtableToParquet"

.github/actions/setup-env/action.yml

@@ -55,12 +55,13 @@ runs:
   using: 'composite'
   steps:
     - name: Setup Go
-      uses: actions/setup-go@44e221478fc6847752e5c574fc7a7b3247b00fbf
+      uses: actions/setup-go@0c52d547c9bc32b1aa3301fd7a9cb496313a4491 # v5.0.0
       with:
         go-version: ${{ inputs.go-version }}
+        cache: false
     - name: Get Changed Files
       id: changed-files
-      uses: tj-actions/changed-files@61ee456a9d0f512e7ecfdf28863634c97dae2d16
+      uses: tj-actions/changed-files@4c5f5d698fbf2d763d5f13815ac7c2ccbef1ff7f # v44.2.0
       with:
         separator: ','
     # It shouldn't take too long to build all of this, and it will at least

.github/actions/setup-java-env/action.yml

@@ -40,7 +40,7 @@ runs:
   using: 'composite'
   steps:
     - name: Setup Java
-      uses: actions/setup-java@a12e082d834968c1847f782019214fadd20719f6
+      uses: actions/setup-java@5896cecc08fd8a1fbdfaf517e29b571164b031f7 # v4.2.0
       with:
         distribution: 'temurin'
         java-version: ${{ inputs.java-version }}
@@ -57,7 +57,7 @@ runs:
           echo "YESTERDAY=$KEY" >> $GITHUB_ENV
         fi
     - name: Setup Cache
-      uses: actions/cache@72d1e4fdff0ff7b1b6e86b415f2d4f5941e5c006
+      uses: actions/cache@ab5e6d0c87105b4c9c2047343972218f562e4319 # v4.0.1
       id: setup-cache
       with:
         path: |

.github/scripts/configure-runners.sh

@@ -13,7 +13,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# For running the script, see go/templates-gitactions-script
+# For running the script, see
+# https://github.com/GoogleCloudPlatform/DataflowTemplates/blob/main/contributor-docs/maintainers-guide.md#provision-new-runners
 
 # Defaults
 NAME_SUFFIX="it"
@@ -22,6 +23,10 @@ BASE_NAME="gitactions-runner"
 REPO_NAME="DataflowTemplates"
 REPO_OWNER="GoogleCloudPlatform"
 GH_RUNNER_VERSION="2.299.1"
+
+MACHINE_TYPE="n1-highmem-32"
+BOOT_DISK_SIZE="200GB"
+
 VERBOSE=0
 
 ############################################################
@@ -44,6 +49,8 @@ Help()
    echo "o     (optional) Set the owner of the GitHub repo. Default '$REPO_OWNER'"
    echo "s     (optional) Set the number of runners. Default $SIZE"
    echo "v     (optional) Set the gitactions runner version. Default $GH_RUNNER_VERSION"
+   echo "m     (optional) Set the machine type for the GCE VM runner. $MACHINE_TYPE"
+   echo "b     (optional) Set the boot disk size for the GCE VM runner. $BOOT_DISK_SIZE"
    echo "V     Verbose mode."
    echo "h     Print this Help."
    echo
@@ -79,6 +86,10 @@ while getopts ":h:Vp:a:t:n:S:r:o:s:v:" option; do
          SIZE=$OPTARG;;
       v) # Enter a version
          GH_RUNNER_VERSION=$OPTARG;;
+      m) # Enter a machine type
+         MACHINE_TYPE=$OPTARG;;
+      b) # Enter a boot disk size
+         BOOT_DISK_SIZE=$OPTARG;;
       V) # Verbose
          VERBOSE=1;;
       \?) # Invalid option
@@ -163,8 +174,6 @@ gcloud secrets add-iam-policy-binding $SECRET_NAME \
 IMAGE_FAMILY="ubuntu-2004-lts"
 IMAGE_PROJECT="ubuntu-os-cloud"
 BOOT_DISK_TYPE="pd-balanced"
-BOOT_DISK_SIZE="200GB"
-MACHINE_TYPE="n1-highmem-16"
 SCOPE="cloud-platform"
 if [ $VERBOSE -eq 1 ]; then echo; echo "Creating instance template: $INSTANCE_TEMPLATE_NAME..."; fi
 if [ $VERBOSE -eq 1 ]; then
@@ -181,7 +190,7 @@ gcloud compute instance-templates create $INSTANCE_TEMPLATE_NAME \
   --image-project=$IMAGE_PROJECT \
   --boot-disk-type=$BOOT_DISK_TYPE \
   --boot-disk-size=$BOOT_DISK_SIZE \
-  --machine-type="MACHINE_TYPE" \
+  --machine-type=$MACHINE_TYPE \
   --scopes=$SCOPE \
   --service-account=${SA_EMAIL} \
   --metadata-from-file=startup-script=startup-script-${NAME_SUFFIX}.sh,shutdown-script=shutdown-script-${NAME_SUFFIX}.sh

.github/scripts/startup-script.sh

@@ -22,30 +22,28 @@ ulimit -n 65536
 # increase max virtual memory
 sudo sysctl -w vm.max_map_count=262144
 
+# update git
+sudo add-apt-repository ppa:git-core/ppa -y
+sudo apt update
+sudo apt install git -y
+
 # install jq
-apt-get update
-apt-get -y install jq
+sudo apt install jq -y
 
 # install maven
-sudo apt update
 sudo apt install git maven -y
 
-# update git
-sudo add-apt-repository ppa:git-core/ppa -y
-sudo apt-get update
-sudo apt-get install git -y
-
 # install gh
-sudo type -p curl >/dev/null || (sudo apt update && sudo apt install curl -y)
-sudo curl -fsSL https://cli.github.com/packages/githubcli-archive-keyring.gpg | sudo dd of=/usr/share/keyrings/githubcli-archive-keyring.gpg \
+sudo apt install curl -y \
+&& sudo curl -fsSL https://cli.github.com/packages/githubcli-archive-keyring.gpg | sudo dd of=/usr/share/keyrings/githubcli-archive-keyring.gpg \
 && sudo chmod go+r /usr/share/keyrings/githubcli-archive-keyring.gpg \
 && echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/githubcli-archive-keyring.gpg] https://cli.github.com/packages stable main" | sudo tee /etc/apt/sources.list.d/github-cli.list > /dev/null \
 && sudo apt update \
 && sudo apt install gh -y
 
 # install docker
-sudo apt-get update
-sudo apt-get install \
+sudo apt update
+sudo apt install \
     ca-certificates \
     curl \
     gnupg \
@@ -55,8 +53,8 @@ curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o
 echo \
   "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu \
   $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
-sudo apt-get update
-sudo apt-get install docker-ce docker-ce-cli containerd.io docker-compose-plugin -y
+sudo apt update
+sudo apt install docker-ce docker-ce-cli containerd.io docker-compose-plugin -y
 
 # add user to docker group
 sudo groupadd docker

.github/workflows/go-pr.yml

@@ -35,9 +35,9 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout Code
-        uses: actions/checkout@ee0669bd1cc54295c223e0bb666b733df41de1c5 # v2.7.0
+        uses: actions/checkout@1e31de5234b9f8995739874a8ce0492dc87873e2 # v4.0.0
       - name: Setup Go
-        uses: actions/setup-go@44e221478fc6847752e5c574fc7a7b3247b00fbf
+        uses: actions/setup-go@0c52d547c9bc32b1aa3301fd7a9cb496313a4491 # v5.0.0
         with:
           go-version: '1.21'
       - name: Run Fmt
@@ -53,9 +53,9 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout Code
-        uses: actions/checkout@ee0669bd1cc54295c223e0bb666b733df41de1c5 # v2.7.0
+        uses: actions/checkout@1e31de5234b9f8995739874a8ce0492dc87873e2 # v4.0.0
       - name: Setup Go
-        uses: actions/setup-go@44e221478fc6847752e5c574fc7a7b3247b00fbf
+        uses: actions/setup-go@0c52d547c9bc32b1aa3301fd7a9cb496313a4491 # v5.0.0
         with:
           go-version: '1.21'
       # By nature, this also makes sure that everything builds

.github/workflows/java-pr.yml

@@ -31,6 +31,10 @@ on:
     - cron: "0 */12 * * *"
   workflow_dispatch:
 
+concurrency:
+  group: java-pr-${{ github.event.issue.number || github.run_id }}
+  cancel-in-progress: true
+
 env:
   MAVEN_OPTS: -Dorg.slf4j.simpleLogger.log.org.apache.maven.plugins.shade=error
 

.github/workflows/prepare-java-cache.yml

@@ -58,7 +58,7 @@ jobs:
         run: |
           echo "CACHE_KEY=''" >> $GITHUB_ENV
       - name: Checkout Code
-        uses: actions/checkout@ee0669bd1cc54295c223e0bb666b733df41de1c5 # v2.7.0
+        uses: actions/checkout@1e31de5234b9f8995739874a8ce0492dc87873e2 # v4.0.0
       - name: Setup Java
         id: setup-java
         uses: ./.github/actions/setup-java-env

.github/workflows/release.yml

@@ -23,7 +23,7 @@ permissions:
 jobs:
   release:
     name: Create Release
-    runs-on: [self-hosted, it]
+    runs-on: [self-hosted, release]
     steps:
     - name: Get releaser identity
       run: |

.github/workflows/scorecards-analysis.yml

@@ -22,7 +22,7 @@ jobs:
 
     steps:
       - name: "Checkout code"
-        uses: actions/checkout@ee0669bd1cc54295c223e0bb666b733df41de1c5 # v2.7.0
+        uses: actions/checkout@1e31de5234b9f8995739874a8ce0492dc87873e2 # v4.0.0
         with:
           persist-credentials: false
 

contributor-docs/maintainers-guide.md

@@ -56,4 +56,73 @@ This repo's code currently is mirrored in Google's internal source control syste
 6) Approve the change internally. At this point, the change will get automatically merged internally and externally.
 
 We are actively working to deprecate this process and use GitHub as the only source of truth.
-If you encounter unresolvable issues with this flow, please reach out to the Dataflow team directly.
+If you encounter unresolvable issues with this flow, please reach out to the Dataflow team directly.
+
+## GitHub actions
+
+There are several workflows that leverage GitHub actions to keep the repo healthy. Of these workflows, there are 
+currently 2 that are run on self-hosted runners on GCP - [Java PR](../.github/workflows/java-pr.yml) which is used to 
+test PR's and [Release](../.github/workflows/release.yml) which is the workflow used for releasing new templates each 
+week.
+
+### Provision new runners
+
+There are instances where we may need to re-provision self-hosted runners, due to unexpected failures, updating 
+dependencies, increasing memory, etc. In these cases, there are helper scripts to aid in redeployment of the GitHub 
+actions runners.
+
+There are 3 scripts: [configure-runners.sh](../.github/scripts/configure-runners.sh), 
+[startup-script.sh](../.github/scripts/startup-script.sh) and 
+[shutdown-script.sh](../.github/scripts/shutdown-script.sh). The first is the main script used to provision the runners 
+themselves. The startup script is what will be invoked by the GCE VM as it is booted up for the first time and will 
+install all necessary packages needed by IT's, unit tests, Release, etc. as well as link the machine as a runner for the 
+repo. Likewise, the shutdown script is run when the VM is shutdown.
+
+To provision GitHub actions runners, there are a couple prerequisites
+- Must be a maintainer of the repo
+- Must have access to GCP project cloud-teleport-testing
+
+Things to remember:
+- Running the script will tear down existing runners and provision new ones. This will kill any actions currently
+running on any of the runners. Failure to spin up new runner correctly will block PR's and Releases, so use carefully.
+- After running the script, it is likely the old runners will still be listed under
+https://github.com/GoogleCloudPlatform/DataflowTemplates/settings/actions/runners. Simply force remove these to keep the
+repo clean
+- The commands below will demonstrate how to provision runners for use with our workflows as they exist today. If there
+arises a need to provision runners in a different manner, feel free to modify the scripts directly and open a PR with 
+the necessary changes.
+
+To run the configuration script:
+
+1. Set gcloud project to cloud-teleport-testing if not already set
+    ```
+    gcloud config set project cloud-teleport-testing
+    ```
+
+2. Export the GitHub actions token
+    ```
+    GITACTIONS_TOKEN=$(gcloud secrets versions access latest --secret=gitactions-runner-secret)
+    ```
+
+3. Run the script
+
+   * For IT runners:
+
+      ```
+      ./configure-runners.sh \
+        -p cloud-teleport-testing \
+        -a 269744978479-compute@developer.gserviceaccount.com \
+        -t $GITACTIONS_TOKEN
+      ```
+
+   * For Performance Test Runner
+      ```
+      ./configure-runners.sh \
+        -p cloud-teleport-testing \
+        -a 269744978479-compute@developer.gserviceaccount.com \
+        -t $GITACTIONS_TOKEN \
+        -S perf \
+        -s 1
+      ```
+
+**Note**: To see optional configurable parameters, run `./configure-runners.sh -h`

it/google-cloud-platform/src/main/java/org/apache/beam/it/gcp/bigquery/BigQueryResourceManagerUtils.java

@@ -22,17 +22,18 @@
 import com.google.cloud.bigquery.TableId;
 import java.time.format.DateTimeFormatter;
 import java.util.regex.Pattern;
+import org.apache.commons.lang3.RandomStringUtils;
 
 /** Utilities for {@link BigQueryResourceManager} implementations. */
 public final class BigQueryResourceManagerUtils {
 
   private static final int MAX_DATASET_ID_LENGTH = 1024;
   private static final Pattern ILLEGAL_DATASET_ID_CHARS = Pattern.compile("[^a-zA-Z0-9_]");
+  private static final String REPLACE_CHAR = "_";
   private static final int MIN_TABLE_ID_LENGTH = 1;
   private static final int MAX_TABLE_ID_LENGTH = 1024;
   private static final Pattern ILLEGAL_TABLE_CHARS = Pattern.compile("[^a-zA-Z0-9-_]");
-  private static final DateTimeFormatter TIME_FORMAT =
-      DateTimeFormatter.ofPattern("yyyyMMdd_HHmmss_SSSSSS");
+  private static final String TIME_FORMAT = "yyyyMMdd_HHmmss";
 
   private BigQueryResourceManagerUtils() {}
 
@@ -46,8 +47,31 @@ private BigQueryResourceManagerUtils() {}
    * @return a BigQuery compatible dataset name.
    */
   static String generateDatasetId(String datasetName) {
+
+    // Take substring of datasetName to account for random suffix
+    // TODO(polber) - remove with Beam 2.57.0
+    int randomSuffixLength = 6;
+    datasetName =
+        datasetName.substring(
+            0,
+            Math.min(
+                datasetName.length(),
+                MAX_DATASET_ID_LENGTH
+                    - REPLACE_CHAR.length()
+                    - TIME_FORMAT.length()
+                    - REPLACE_CHAR.length()
+                    - randomSuffixLength));
+
+    // Add random suffix to avoid collision
+    // TODO(polber) - remove with Beam 2.57.0
     return generateResourceId(
-        datasetName, ILLEGAL_DATASET_ID_CHARS, "_", MAX_DATASET_ID_LENGTH, TIME_FORMAT);
+            datasetName,
+            ILLEGAL_DATASET_ID_CHARS,
+            REPLACE_CHAR,
+            MAX_DATASET_ID_LENGTH,
+            DateTimeFormatter.ofPattern(TIME_FORMAT))
+        + REPLACE_CHAR
+        + RandomStringUtils.randomAlphanumeric(6).toLowerCase();
   }
 
   /**