diff --git a/.github/workflows/ci_test-mnodes.yml b/.github/workflows/ci_test-mnodes.yml
new file mode 100644
index 0000000000..259ce9642e
--- /dev/null
+++ b/.github/workflows/ci_test-mnodes.yml
@@ -0,0 +1,212 @@
+name: Multi Nodes GPU Tests
+
+# Workflow Steps:
+#  1. Checkout Pytorch Lightning
+#  2. Set up Python
+#  3. Configure AWS Credentials
+#  4. Install AWS Client
+#  5. Get Current Sha Commit
+#  6. Create Job Name
+#  7. Update Test Configuration File
+#  8. Install EKSClient
+#  9. Create Gpu Node Pool
+#  10. Check Current Node Pool | Current Elatic Pods
+#  11. Apply Elastic
+#  12. Wait 5 sec
+#  13. Find ETCD TCP Address
+#  14. Update Test Configuration File
+#  15. Apply Multi Node Testing
+#  16. Wait 120 secs
+#  17. Listen to Jobs Logging
+#  18. Statistics
+#  19. Upload coverage results
+#  20. Upload coverage to Codecov
+#  21. Delete Group Node
+
+#on: push
+
+on:
+  push:
+    branches:
+      - master
+      - release/*
+  pull_request:
+    types: [closed]
+
+env:
+  AWS_CLUSTER: pl-lightning-torchelastic
+  NODE_TYPE: g4dn.xlarge
+  NODES: 2
+  NUM_GPUS: 1
+  REGION: us-east-2
+  MAX_CHECKS: 300
+  CHECK_SPEEP: 2
+
+jobs:
+  multi-nodes-gpu-testing:
+    runs-on: ubuntu-20.04
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: [3.7]
+        pytorch-version: [1.5]
+    # Timeout: https://stackoverflow.com/a/59076067/4521646
+    timeout-minutes: 50
+
+    # runs only when merged happened.
+    # if: github.event.pull_request.merged == true
+    steps:
+
+    - name: Checkout Pytorch Lightning
+      uses: actions/checkout@v2
+      with:
+        repository: PyTorchLightning/pytorch-lightning
+        ref: ${{ github.event.base_ref }}
+
+    - name: Set up Python
+      uses: actions/setup-python@v2
+      with:
+        python-version: ${{ matrix.python-version }}
+
+    # Note: This uses an internal pip API and may not always work
+    # https://github.com/actions/cache/blob/master/examples.md#multiple-oss-in-a-workflow
+    - name: Cache pip
+      uses: actions/cache@v2
+      with:
+        path: ~/.cache/pip
+        key: ${{ runner.os }}-pip-multi-node
+        restore-keys: |
+          ${{ runner.os }}-pip-
+
+    - name: Install dependencies
+      run: |
+        pip install awscli coverage
+        # todo
+        pip install git+https://${{ secrets.PL_GHOST_TOKEN }}@github.com/PyTorchLightning/lightning-dtrun.git@v0.0.3 -q --no-cache-dir
+        #pip install git+https://${{ secrets.PL_GHOST_TOKEN }}@github.com/PyTorchLightning/lightning-dtrun.git@mnodes -q --no-cache-dir
+
+    - name: Configure AWS Credentials
+      uses: aws-actions/configure-aws-credentials@v1
+      with:
+        aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
+        aws-secret-access-key: ${{ secrets.AWS_SECRET_KEY_ID }}
+        aws-region: us-east-2
+
+    - name: Get Current Sha Commit
+      id: vars
+      shell: bash
+      run: |
+        echo "::set-output name=SHA::$(git rev-parse --short HEAD)"
+        echo $PWD
+
+    - name: Create Job Name
+      id: job
+      shell: bash
+      run: |
+        echo "::set-output name=ID::$(echo '${{ steps.vars.outputs.SHA }}-${{ matrix.python-version }}-${{ matrix.pytorch-version }}' | tr . - )"
+        echo "::set-output name=ID_NAME::$(echo 's-${{ steps.vars.outputs.SHA }}-${{ matrix.python-version }}-${{ matrix.pytorch-version }}-e' | tr . - )"
+
+    - name: Install EKSClient
+      run: |
+        curl --silent --location "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" | tar xz -C /tmp
+        sudo mv /tmp/eksctl /usr/local/bin
+      shell: bash
+
+    - name: Create Gpu Node Pool
+      run: |
+        aws eks --region $REGION update-kubeconfig --name $AWS_CLUSTER
+        eksctl create nodegroup --name=${{ steps.job.outputs.ID }} --cluster=$AWS_CLUSTER --node-type=$NODE_TYPE  --nodes=$NODES
+        # eksctl create nodegroup --name=${{ steps.job.outputs.ID }} --cluster=$AWS_CLUSTER --managed --spot --node-type=$NODE_TYPE  --nodes=$NODES
+      shell: bash
+
+    - name: Check Current Node Pool | Current Elatic Pods
+      run: |
+        eksctl get nodegroups --cluster $AWS_CLUSTER
+        kubectl get pods -n elastic-job
+
+    - name: Apply Elastic
+      run: |
+        git clone https://github.com/pytorch/elastic.git
+        cd elastic/kubernetes
+
+        kubectl apply -k config/default
+
+        kubectl apply -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/master/nvidia-device-plugin.yml
+        kubectl apply -f https://raw.githubusercontent.com/pytorch/elastic/master/kubernetes/config/samples/etcd.yaml
+
+    - name: Wait
+      # todo: this shall be dynamic
+      if: always()
+      shell: bash
+      run: |
+        sleep 5
+
+    - name: Find ETCD TCP Address
+      id: tcp
+      shell: bash
+      run: |
+        echo "::set-output name=TCP_ADDRESS::$(kubectl logs etcd -n elastic-job | grep -Eo '[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}:[0-9]{1,4}' | head -1)"
+
+    - name: Update Test Config. File
+      run: |
+        import os
+        from dtrun.configs import prepare_multi_nodes_gpu_config
+
+        assert os.path.isfile('./tests/mnode_tests.txt')
+        prepare_multi_nodes_gpu_config(
+            './.github/multi-nodes-gpu.yaml',
+            './tests/mnode_tests.txt',
+            sha="${{ steps.vars.outputs.SHA }}",
+            tcp_address="${{ steps.tcp.outputs.TCP_ADDRESS }}",
+            python_version="${{ matrix.python-version }}",
+            torch_version="${{ matrix.pytorch-version }}",
+            num_gpus=1,
+        )
+      shell: python
+
+    - name: Apply Multi Node Testing
+      run: |
+        # cat ./.github/multi-nodes-gpu.yaml
+        kubectl apply -f ./.github/multi-nodes-gpu.yaml
+      shell: bash
+
+    - name: Wait
+      # todo: this shall be dynamic
+      if: always()
+      shell: bash
+      run: |
+        sleep 400
+
+    - name: Listen to Jobs Logging
+      shell: bash
+      run: |
+        # todo: Enable automatic checking.
+        # while [ $i -lt $MAX_CHECKS ]; do ((i++)); if kubectl logs ${{ steps.job.outputs.ID_NAME }}-worker-0 -n elastic-job | grep -i "error\|failed"; then status_code=1 && break; elif kubectl logs ${{ steps.job.outputs.ID }}-worker-0 -n elastic-job | grep "TEST END"; then status_code=0 && break; else printf "." ; fi; sleep $CHECK_SPEEP; done && \
+        # echo "Done waiting. Job status code: $status_code" && \
+        kubectl logs ${{ steps.job.outputs.ID_NAME }}-worker-0 -n elastic-job > /tmp/full_output.txt
+        if grep -q 'END_TOKEN' /tmp/full_output.txt ; then csplit /tmp/full_output.txt '/END_TOKEN/'; else mv /tmp/full_output.txt xx00; fi && \
+        cat xx00
+
+    - name: Statistics
+      if: success()
+      run: |
+        cat ./xx01  | tail -n +2 | base64 --decode > /home/runner/work/pytorch-lightning/pytorch-lightning/.coverage
+        cd /home/runner/work/pytorch-lightning/pytorch-lightning && coverage report && coverage xml
+
+    - name: Upload coverage to Codecov
+      uses: codecov/codecov-action@v1
+      if: always()
+      # see: https://github.com/actions/toolkit/issues/399
+      continue-on-error: true
+      with:
+        token: ${{ secrets.CODECOV_TOKEN }}
+        file: coverage.xml
+        flags: multi-nodes,pytest
+        name: multi-nodes-coverage
+        fail_ci_if_error: false
+
+    - name: Delete Group Node
+      if: always()
+      run: |
+       kubectl delete  ElasticJob ${{ steps.job.outputs.ID_NAME }} -n elastic-job
+       eksctl delete nodegroup ${{ steps.job.outputs.ID }} --cluster=$AWS_CLUSTER
diff --git a/tests/accelerators/legacy/test_multi_nodes_gpu.py b/tests/accelerators/legacy/test_multi_nodes_gpu.py
new file mode 100644
index 0000000000..f17ac42fcb
--- /dev/null
+++ b/tests/accelerators/legacy/test_multi_nodes_gpu.py
@@ -0,0 +1,126 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import os
+import sys
+
+import pytest
+import torch
+
+ROOT = os.path.join(os.path.dirname(os.path.realpath(__file__)), "..")
+sys.path.insert(0, ROOT)
+DIR_PATH = os.path.dirname(os.path.realpath(__file__))
+
+from pytorch_lightning import LightningModule  # noqa: E402
+from pytorch_lightning import Trainer
+from tests.base.boring_model import BoringModel  # noqa: E402
+
+
+@pytest.mark.skipif(not os.getenv("PL_RUNNING_SPECIAL_TESTS", '0') == '1', reason="test should be run outside of pytest")
+def test_logging_sync_dist_true_ddp(tmpdir):
+    """
+    Tests to ensure that the sync_dist flag works with CPU (should just return the original value)
+    """
+    fake_result = 1
+
+    class TestModel(BoringModel):
+        def training_step(self, batch, batch_idx):
+            acc = self.step(batch[0])
+            self.log('foo', torch.tensor(fake_result), on_step=False, on_epoch=True)
+            return acc
+
+        def validation_step(self, batch, batch_idx):
+            output = self.layer(batch)
+            loss = self.loss(batch, output)
+            self.log('bar', torch.tensor(fake_result), on_step=False, on_epoch=True)
+            return {"x": loss}
+
+    model = TestModel()
+    trainer = Trainer(
+        default_root_dir=tmpdir,
+        limit_train_batches=1,
+        limit_val_batches=1,
+        max_epochs=2,
+        weights_summary=None,
+        accelerator="ddp",
+        gpus=1,
+        num_nodes=2,
+    )
+    trainer.fit(model)
+
+    assert trainer.logged_metrics['foo'] == fake_result
+    assert trainer.logged_metrics['bar'] == fake_result
+
+
+@pytest.mark.skipif(not os.getenv("PL_RUNNING_SPECIAL_TESTS", '0') == '1', reason="test should be run outside of pytest")
+def test__validation_step__log(tmpdir):
+    """
+    Tests that validation_step can log
+    """
+    os.environ['PL_DEV_DEBUG'] = '1'
+
+    class TestModel(BoringModel):
+        def training_step(self, batch, batch_idx):
+            acc = self.step(batch)
+            acc = acc + batch_idx
+            self.log('a', acc, on_step=True, on_epoch=True)
+            self.log('a2', 2)
+
+            self.training_step_called = True
+            return acc
+
+        def validation_step(self, batch, batch_idx):
+            acc = self.step(batch)
+            acc = acc + batch_idx
+            self.log('b', acc, on_step=True, on_epoch=True)
+            self.training_step_called = True
+
+        def backward(self, loss, optimizer, optimizer_idx):
+            return LightningModule.backward(self, loss, optimizer, optimizer_idx)
+
+    model = TestModel()
+    model.validation_step_end = None
+    model.validation_epoch_end = None
+
+    trainer = Trainer(
+        default_root_dir=tmpdir,
+        limit_train_batches=2,
+        limit_val_batches=2,
+        max_epochs=2,
+        log_every_n_steps=1,
+        weights_summary=None,
+        accelerator="ddp",
+        gpus=1,
+        num_nodes=2,
+    )
+    trainer.fit(model)
+
+    # make sure all the metrics are available for callbacks
+    expected_logged_metrics = {
+        'a2',
+        'a_step',
+        'a_epoch',
+        'b_step/epoch_0',
+        'b_step/epoch_1',
+        'b_epoch',
+        'epoch',
+    }
+    logged_metrics = set(trainer.logged_metrics.keys())
+    assert expected_logged_metrics == logged_metrics
+
+    # we don't want to enable val metrics during steps because it is not something that users should do
+    # on purpose DO NOT allow step_b... it's silly to monitor val step metrics
+    callback_metrics = set(trainer.callback_metrics.keys())
+    callback_metrics.remove('debug_epoch')
+    expected_cb_metrics = {'a', 'a2', 'b', 'a_epoch', 'b_epoch', 'a_step'}
+    assert expected_cb_metrics == callback_metrics
diff --git a/tests/mnode_tests.txt b/tests/mnode_tests.txt
new file mode 100644
index 0000000000..77a3ed58db
--- /dev/null
+++ b/tests/mnode_tests.txt
@@ -0,0 +1,2 @@
+./tests/backends/test_multi_nodes_gpu.py::test_logging_sync_dist_true_ddp
+./tests/backends/test_multi_nodes_gpu.py::test__validation_step__log