fix(README): resolving a merge conflict

Signed-off-by: Cryptophobia <aouzounov@gmail.com>

README.md

@@ -4,11 +4,10 @@
 [![Build Status](https://travis-ci.org/teamhephy/controller.svg?branch=master)](https://travis-ci.org/teamhephy/controller)
 [![codecov.io](https://codecov.io/github/deis/controller/coverage.svg?branch=master)](https://codecov.io/github/deis/controller?branch=master)
 [![Docker Repository on Quay](https://quay.io/repository/deisci/controller/status "Docker Repository on Quay")](https://quay.io/repository/deisci/controller)
-[![Dependency Status](https://www.versioneye.com/user/projects/5863f1de6f4bf900128fa95a/badge.svg?style=flat)](https://www.versioneye.com/user/projects/5863f1de6f4bf900128fa95a)
 
 Deis (pronounced DAY-iss) Workflow is an open source Platform as a Service (PaaS) that adds a developer-friendly layer to any [Kubernetes](http://kubernetes.io) cluster, making it easy to deploy and manage applications on your own servers.
 
-For more information about the Deis Workflow, please visit the main project page at https://github.com/deisthree/workflow.
+For more information about the Deis Workflow, please visit the main project page at https://github.com/teamhephy/workflow.
 
 We welcome your input! If you have feedback, please [submit an issue][issues]. If you'd like to participate in development, please read the "Development" section below and [submit a pull request][prs].
 
@@ -47,7 +46,7 @@ You'll want to test your code changes interactively in a working Kubernetes clus
 
 ### Workflow Installation
 
-After you have a working Kubernetes cluster, you're ready to [install Workflow](https://deis.com/docs/workflow/installing-workflow/).
+After you have a working Kubernetes cluster, you're ready to [install Workflow](https://docs.teamhephy.com/installing-workflow/).
 
 ## Testing Your Code
 
@@ -77,8 +76,8 @@ kubectl get pod --namespace=deis -w | grep deis-controller
 ```
 
 [install-k8s]: https://kubernetes.io/docs/setup/pick-right-solution
-[issues]: https://github.com/deisthree/controller/issues
-[prs]: https://github.com/deisthree/controller/pulls
-[workflow]: https://github.com/deisthree/workflow
+[issues]: https://github.com/teamhephy/controller/issues
+[prs]: https://github.com/teamhephy/controller/pulls
+[workflow]: https://github.com/teamhephy/workflow
 [Docker]: https://www.docker.com/
-[v2.18]: https://github.com/deisthree/workflow/releases/tag/v2.18.0
+[v2.18]: https://github.com/teamhephy/workflow/releases/tag/v2.21.4

charts/controller/templates/controller-deployment.yaml

@@ -134,6 +134,10 @@ spec:
                 secretKeyRef:
                   name: database-creds
                   key: password
+{{- if (.Values.deis_ignore_scheduling_failure) }}
+            - name: DEIS_IGNORE_SCHEDULING_FAILURE
+              value: "{{ .Values.deis_ignore_scheduling_failure }}"
+{{- end }}
             - name: RESERVED_NAMES
               value: "deis, deis-builder, deis-workflow-manager, grafana"
             - name: WORKFLOW_NAMESPACE

rootfs/Dockerfile

@@ -1,4 +1,4 @@
-FROM quay.io/deis/base:v0.3.6
+FROM hephy/base:v0.4.1
 
 RUN adduser --system \
 	--shell /bin/bash \
@@ -17,6 +17,7 @@ RUN buildDeps='gcc libffi-dev libpq-dev libldap2-dev libsasl2-dev python3-dev py
         libpq5 \
         libldap-2.4 \
         python3-minimal \
+        python3-distutils \
         # cryptography package needs pkg_resources
         python3-pkg-resources && \
     ln -s /usr/bin/python3 /usr/bin/python && \

rootfs/Dockerfile.test

@@ -1,4 +1,4 @@
-FROM quay.io/deis/base:v0.3.6
+FROM hephy/base:v0.4.1
 
 RUN adduser --system \
 	--shell /bin/bash \
@@ -49,7 +49,7 @@ RUN buildDeps='gcc libffi-dev libpq-dev libldap2-dev libsasl2-dev python3-dev py
 WORKDIR /app
 
 # test-unit additions to the main Dockerfile
-ENV PGBIN=/usr/lib/postgresql/9.5/bin PGDATA=/var/lib/postgresql/data
+ENV PGBIN=/usr/lib/postgresql/10/bin PGDATA=/var/lib/postgresql/data
 RUN apt-get update && \
     DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
         git \

rootfs/api/settings/production.py

@@ -218,6 +218,12 @@
             'filters': ['require_debug_true'],
             'propagate': True,
         },
+        'django_auth_ldap': {
+            'handlers': ['console'],
+            'level': 'DEBUG',
+            'filters': ['require_debug_true'],
+            'propagate': False,
+        },
         'api': {
             'handlers': ['console'],
             'propagate': True,

rootfs/requirements.txt

@@ -1,6 +1,6 @@
 # Deis controller requirements
 backoff==1.4.3
-django==1.11.23
+django==1.11.29
 django-auth-ldap==1.2.15
 django-cors-middleware==1.3.1
 django-guardian==1.4.9

rootfs/scheduler/resources/deployment.py

@@ -139,6 +139,11 @@ def create(self, namespace, name, image, entrypoint, command, spec_annotations,
         return response
 
     def update(self, namespace, name, image, entrypoint, command, spec_annotations, **kwargs):
+        # Set the replicas value to the current replicas of the deployment.
+        # This avoids resetting the replicas which causes disruptions during the deployment.
+        deployment = self.deployment.get(namespace, name).json()
+        current_replicas = int(deployment['spec']['replicas'])
+        kwargs['replicas'] = current_replicas
         manifest = self.manifest(namespace, name, image,
                                  entrypoint, command, spec_annotations, **kwargs)
 
@@ -277,9 +282,9 @@ def are_replicas_ready(self, namespace, name):
 
         if (
             'unavailableReplicas' in status or
-            ('replicas' not in status or status['replicas'] is not desired) or
-            ('updatedReplicas' not in status or status['updatedReplicas'] is not desired) or
-            ('availableReplicas' not in status or status['availableReplicas'] is not desired)
+            ('replicas' not in status or status['replicas'] != desired) or
+            ('updatedReplicas' not in status or status['updatedReplicas'] != desired) or
+            ('availableReplicas' not in status or status['availableReplicas'] != desired)
         ):
             return False, pods
 
@@ -380,22 +385,36 @@ def _check_for_failed_events(self, namespace, labels):
         Request for new ReplicaSet of Deployment and search for failed events involved by that RS
         Raises: KubeException when RS have events with FailedCreate reason
         """
-        response = self.rs.get(namespace, labels=labels)
-        data = response.json()
-        fields = {
-            'involvedObject.kind': 'ReplicaSet',
-            'involvedObject.name': data['items'][0]['metadata']['name'],
-            'involvedObject.namespace': namespace,
-            'involvedObject.uid': data['items'][0]['metadata']['uid'],
-        }
-        events_list = self.ns.events(namespace, fields=fields).json()
-        events = events_list.get('items', [])
-        if events is not None and len(events) != 0:
-            for event in events:
-                if event['reason'] == 'FailedCreate':
-                    log = self._get_formatted_messages(events)
-                    self.log(namespace, log)
-                    raise KubeException(log)
+        max_retries = 3
+        retry_sleep_sec = 3.0
+        for try_ in range(max_retries):
+            response = self.rs.get(namespace, labels=labels)
+            data = response.json()
+            try:
+                fields = {
+                    'involvedObject.kind': 'ReplicaSet',
+                    'involvedObject.name': data['items'][0]['metadata']['name'],
+                    'involvedObject.namespace': namespace,
+                    'involvedObject.uid': data['items'][0]['metadata']['uid'],
+                }
+            except Exception as e:
+                if try_ + 1 < max_retries:
+                    self.log(namespace,
+                             "Got an empty ReplicaSet list. Trying one more time. {}".format(
+                                 json.dumps(labels)))
+                    time.sleep(retry_sleep_sec)
+                    continue
+                self.log(namespace, "Did not find the ReplicaSet for {}".format(
+                    json.dumps(labels)), "WARN")
+                raise e
+            events_list = self.ns.events(namespace, fields=fields).json()
+            events = events_list.get('items', [])
+            if events is not None and len(events) != 0:
+                for event in events:
+                    if event['reason'] == 'FailedCreate':
+                        log = self._get_formatted_messages(events)
+                        self.log(namespace, log)
+                        raise KubeException(log)
 
     @staticmethod
     def _get_formatted_messages(events):

rootfs/scheduler/resources/pod.py

@@ -552,7 +552,7 @@ def events(self, pod):
         if not events:
             events = []
         # make sure that events are sorted
-        events.sort(key=lambda x: x['lastTimestamp'])
+        events.sort(key=lambda x: x['lastTimestamp'] or '')
         return events
 
     def _handle_pod_errors(self, pod, reason, message):
@@ -577,9 +577,11 @@ def _handle_pod_errors(self, pod, reason, message):
             "ErrImageNeverPull": "ErrImageNeverPullPolicy",
             # Not including this one for now as the message is not useful
             # "BackOff": "BackOffPullImage",
-            # FailedScheduling relates limits
-            "FailedScheduling": "FailedScheduling",
         }
+        # We want to be able to ignore pod scheduling errors as they might be temporary
+        if not os.environ.get("DEIS_IGNORE_SCHEDULING_FAILURE", False):
+            # FailedScheduling relates limits
+            event_errors["FailedScheduling"] = "FailedScheduling"
 
         # Nicer error than from the event
         # Often this gets to ImageBullBackOff before we can introspect tho

rootfs/scheduler/tests/test_deployments.py

@@ -102,12 +102,12 @@ def test_deployment_api_version_1_9_and_up(self):
             deployment.version = mock.MagicMock(return_value=parse(canonical))
             actual = deployment.api_version
             self.assertEqual(
+                expected,
+                actual,
+                "{} breaks - expected {}, got {}".format(
+                    canonical,
                     expected,
-                    actual,
-                    "{} breaks - expected {}, got {}".format(
-                        canonical,
-                        expected,
-                        actual))
+                    actual))
 
     def test_deployment_api_version_1_8_and_lower(self):
         cases = ['1.8', '1.7', '1.6', '1.5', '1.4', '1.3', '1.2']
@@ -120,12 +120,12 @@ def test_deployment_api_version_1_8_and_lower(self):
             deployment.version = mock.MagicMock(return_value=parse(canonical))
             actual = deployment.api_version
             self.assertEqual(
+                expected,
+                actual,
+                "{} breaks - expected {}, got {}".format(
+                    canonical,
                     expected,
-                    actual,
-                    "{} breaks - expected {}, got {}".format(
-                        canonical,
-                        expected,
-                        actual))
+                    actual))
 
     def test_create_failure(self):
         with self.assertRaises(
@@ -158,11 +158,13 @@ def test_update(self):
         deployment = self.scheduler.deployment.get(self.namespace, name).json()
         self.assertEqual(deployment['spec']['replicas'], 4, deployment)
 
-        # emulate scale without calling scale
-        self.update(self.namespace, name, replicas=2)
+        # update the version
+        new_version = 'v1024'
+        self.update(self.namespace, name, version=new_version)
 
         deployment = self.scheduler.deployment.get(self.namespace, name).json()
-        self.assertEqual(deployment['spec']['replicas'], 2, deployment)
+        self.assertEqual(deployment['spec']['template']['metadata']['labels']['version'],
+                         new_version, deployment)
 
     def test_delete_failure(self):
         # test failure

rootfs/scheduler/tests/test_horizontalpodautoscaler.py

@@ -62,27 +62,14 @@ def update(self, namespace=None, name=generate_random_name(), **kwargs):
         self.assertEqual(horizontalpodautoscaler.status_code, 200, horizontalpodautoscaler.json())  # noqa
         return name
 
-    def update_deployment(self, namespace=None, name=generate_random_name(), **kwargs):
+    def scale_deployment(self, namespace=None, name=generate_random_name(), replicas=1):
         """
-        Helper function to update and verify a deployment on the namespace
+        Helper function to scale the replicas of a deployment
         """
         namespace = self.namespace if namespace is None else namespace
-        # these are all required even if it is kwargs...
-        d_kwargs = {
-            'app_type': kwargs.get('app_type', 'web'),
-            'version': kwargs.get('version', 'v99'),
-            'replicas': kwargs.get('replicas', 4),
-            'pod_termination_grace_period_seconds': 2,
-            'image': 'quay.io/fake/image',
-            'entrypoint': 'sh',
-            'command': 'start',
-            'spec_annotations': kwargs.get('spec_annotations', {}),
-        }
-
-        deployment = self.scheduler.deployment.update(namespace, name, **d_kwargs)
-        data = deployment.json()
-        self.assertEqual(deployment.status_code, 200, data)
-        return name
+        self.scheduler.deployment.scale(namespace, name, image=None,
+                                        entrypoint=None, command=None,
+                                        replicas=replicas)
 
     def test_create_failure(self):
         with self.assertRaises(
@@ -147,7 +134,7 @@ def test_update(self):
         self.assertEqual(deployment['status']['availableReplicas'], 3)
 
         # scale deployment to 1 (should go back to 3)
-        self.update_deployment(self.namespace, name, replicas=1)
+        self.scale_deployment(self.namespace, name, replicas=1)
 
         # check the deployment object
         deployment = self.scheduler.deployment.get(self.namespace, name).json()
@@ -158,7 +145,7 @@ def test_update(self):
         self.assertEqual(deployment['status']['availableReplicas'], 3)
 
         # scale deployment to 6 (should go back to 4)
-        self.update_deployment(self.namespace, name, replicas=6)
+        self.scale_deployment(self.namespace, name, replicas=6)
 
         # check the deployment object
         deployment = self.scheduler.deployment.get(self.namespace, name).json()

rootfs/scheduler/tests/test_horizontalpodautoscaler_12_lower.py

@@ -66,27 +66,14 @@ def update(self, namespace=None, name=generate_random_name(), **kwargs):
         self.assertEqual(horizontalpodautoscaler.status_code, 200, horizontalpodautoscaler.json())  # noqa
         return name
 
-    def update_deployment(self, namespace=None, name=generate_random_name(), **kwargs):
+    def scale_deployment(self, namespace=None, name=generate_random_name(), replicas=1):
         """
-        Helper function to update and verify a deployment on the namespace
+        Helper function to scale the replicas of a deployment
         """
         namespace = self.namespace if namespace is None else namespace
-        # these are all required even if it is kwargs...
-        kwargs = {
-            'app_type': kwargs.get('app_type', 'web'),
-            'version': kwargs.get('version', 'v99'),
-            'replicas': kwargs.get('replicas', 4),
-            'pod_termination_grace_period_seconds': 2,
-            'image': 'quay.io/fake/image',
-            'entrypoint': 'sh',
-            'command': 'start',
-            'spec_annotations': kwargs.get('spec_annotations', {}),
-        }
-
-        deployment = self.scheduler.deployment.update(namespace, name, **kwargs)
-        data = deployment.json()
-        self.assertEqual(deployment.status_code, 200, data)
-        return name
+        self.scheduler.deployment.scale(namespace, name, image=None,
+                                        entrypoint=None, command=None,
+                                        replicas=replicas)
 
     def test_create_failure(self):
         with self.assertRaises(
@@ -151,7 +138,7 @@ def test_update(self):
         self.assertEqual(deployment['status']['availableReplicas'], 3)
 
         # scale deployment to 1 (should go back to 3)
-        self.update_deployment(self.namespace, name, replicas=1)
+        self.scale_deployment(self.namespace, name, replicas=1)
 
         # check the deployment object
         deployment = self.scheduler.deployment.get(self.namespace, name).json()
@@ -162,7 +149,7 @@ def test_update(self):
         self.assertEqual(deployment['status']['availableReplicas'], 3)
 
         # scale deployment to 6 (should go back to 4)
-        self.update_deployment(self.namespace, name, replicas=6)
+        self.scale_deployment(self.namespace, name, replicas=6)
 
         # check the deployment object
         deployment = self.scheduler.deployment.get(self.namespace, name).json()