Fix jenkins (#3875)

* try fix * try fix * try fix * try fix * Revert "try fix" This reverts commit a3fa0b1e9c0ab892cc3a22acf3770903db8b14a7. * try fix shared memory * try fix shared memory * try fix image version * fix

Fix jenkins (#3875)
* try fix * try fix * try fix * try fix * Revert "try fix" This reverts commit a3fa0b1e9c0ab892cc3a22acf3770903db8b14a7. * try fix shared memory * try fix shared memory * try fix image version * fix
a3fd0595 · Jinjing Zhou · GitHub · 480a4ae3 · a3fd0595 · a3fd0595
Unverified Commit a3fd0595 authored Mar 23, 2022 by Jinjing Zhou Committed by GitHub Mar 23, 2022
Hide whitespace changes
Inline Side-by-side

Showing with 57 additions and 38 deletions

Jenkinsfile Jenkinsfile +55 -36

tests/pytorch/test_optim.py tests/pytorch/test_optim.py +2 -2

No files found.
--- a/Jenkinsfile
+++ b/Jenkinsfile
@@ -152,9 +152,10 @@ pipeline {
    }
    stage('Bot Instruction') {
      agent {
-        kubernetes {
+        docker {
-          yamlFile 'docker/pods/ci-lint.yaml'
+            label 'linux-benchmark-node'
-          defaultContainer 'dgl-ci-lint'
+            image 'dgllib/dgl-ci-lint'
+            alwaysPull true
        }
      }
      steps {
@@ -174,9 +175,10 @@ pipeline {
      stages {
        stage('Lint Check') {
          agent {
-            kubernetes {
+            docker {
-              yamlFile 'docker/pods/ci-lint.yaml'
+              label "linux-cpu-node"
-              defaultContainer 'dgl-ci-lint'
+              image "dgllib/dgl-ci-lint"  
+              alwaysPull true
            }
          }
          steps {
@@ -194,9 +196,11 @@ pipeline {
          parallel {
            stage('CPU Build') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-compile-cpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-cpu-compile'
+                  image "dgllib/dgl-ci-cpu:cu101_v220217"  
+                  args "-u root"
+                  alwaysPull true
                }
              }
              steps {
@@ -210,9 +214,11 @@ pipeline {
            }
            stage('GPU Build') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-compile-gpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-gpu-compile'
+                  image "dgllib/dgl-ci-gpu:cu101_v220217"  
+                  args "-u root"
+                  alwaysPull true
                }
              }
              steps {
@@ -245,9 +251,10 @@ pipeline {
          parallel {
            stage('C++ CPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-cpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-cpu'
+                  image "dgllib/dgl-ci-cpu:cu101_v220217"  
+                  alwaysPull true
                }
              }
              steps {
@@ -261,9 +268,11 @@ pipeline {
            }
            stage('C++ GPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-gpu.yaml'
+                  label "linux-gpu-node"
-                  defaultContainer 'dgl-ci-gpu'
+                  image "dgllib/dgl-ci-gpu:cu101_v220217"  
+                  args "--runtime nvidia"
+                  alwaysPull true
                }
              }
              steps {
@@ -288,9 +297,10 @@ pipeline {
            }
            stage('Tensorflow CPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-cpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-cpu'
+                  image "dgllib/dgl-ci-cpu:cu101_v220217"  
+                  alwaysPull true
                }
              }
              stages {
@@ -308,9 +318,11 @@ pipeline {
            }
            stage('Tensorflow GPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-gpu.yaml'
+                  label "linux-gpu-node"
-                  defaultContainer 'dgl-ci-gpu'
+                  image "dgllib/dgl-ci-gpu:cu101_v220217"  
+                  args "--runtime nvidia"
+                  alwaysPull true
                }
              }
              stages {
@@ -328,9 +340,11 @@ pipeline {
            }
            stage('Torch CPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-cpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-cpu'
+                  image "dgllib/dgl-ci-cpu:cu101_v220217"  
+                  args "--shm-size=4gb"
+                  alwaysPull true
                }
              }
              stages {
@@ -378,9 +392,11 @@ pipeline {
            }
            stage('Torch GPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-gpu.yaml'
+                  label "linux-gpu-node"
-                  defaultContainer 'dgl-ci-gpu'
+                  image "dgllib/dgl-ci-gpu:cu101_v220217"  
+                  args "--runtime nvidia --shm-size=8gb"
+                  alwaysPull true
                }
              }
              stages {
@@ -404,9 +420,10 @@ pipeline {
            }
            stage('MXNet CPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-cpu.yaml'
+                  label "linux-cpu-node"
-                  defaultContainer 'dgl-ci-cpu'
+                  image "dgllib/dgl-ci-cpu:cu101_v220217"  
+                  alwaysPull true
                }
              }
              stages {
@@ -429,9 +446,11 @@ pipeline {
            }
            stage('MXNet GPU') {
              agent {
-                kubernetes {
+                docker {
-                  yamlFile 'docker/pods/ci-gpu.yaml'
+                  label "linux-gpu-node"
-                  defaultContainer 'dgl-ci-gpu'
+                  image "dgllib/dgl-ci-gpu:cu101_v220217"  
+                  args "--runtime nvidia"
+                  alwaysPull true
                }
              }
              stages {

--- a/tests/pytorch/test_optim.py
+++ b/tests/pytorch/test_optim.py
@@ -178,7 +178,7 @@ def start_torch_adam_worker(rank, world_size, weight, has_zero_grad=False,
    th.distributed.barrier()
 @unittest.skipIf(os.name == 'nt', reason='Do not support windows yet')
-@unittest.skipIf(F.ctx().type == 'gpu', reason='cpu only test')
+@unittest.skipIf(F.ctx().type != 'cpu', reason='cpu only test')
 @pytest.mark.parametrize("num_workers", [2, 4])
 def test_multiprocess_cpu_sparse_adam(num_workers):
    backend = 'gloo'
@@ -282,7 +282,7 @@ def test_multiprocess_sparse_adam_cuda_tensor(num_workers):
    assert F.allclose(dgl_weight, torch_weight)
 @unittest.skipIf(os.name == 'nt', reason='Do not support windows yet')
-@unittest.skipIf(F.ctx().type == 'gpu', reason='cpu only test')
+@unittest.skipIf(F.ctx().type != 'cpu', reason='cpu only test')
 @pytest.mark.parametrize("num_workers", [2, 4])
 def test_multiprocess_sparse_adam_cpu_zero_step(num_workers):
    backend = 'gloo'