Compare commits

..

1 Commits

Author SHA1 Message Date
David O Neill
8369c43123 Stale action for prs/issues 2024-01-30 16:10:13 +00:00
136 changed files with 838 additions and 2679 deletions

31
.github/actions/stale/stale.yml vendored Normal file
View File

@@ -0,0 +1,31 @@
name: 'Close stale issues and PRs'
on:
schedule:
- cron: '30 1 * * *'
jobs:
stale:
runs-on: ubuntu-latest
steps:
- uses: actions/stale@v9
with:
debug-only: true
operations-per-run: 30
days-before-stale: 30
days-before-close: 5
days-before-issue-stale: 180
days-before-issue-close: 14
stale-issue-message: 'This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.'
close-issue-message: 'This issue was closed because it has been stalled for 5 days with no activity.'
exempt-issue-labels: 'needs_triage'
stale-issue-label: 'no_issue_activity'
days-before-pr-stale: 90
days-before-pr-close: 14
stale-pr-message: 'This PR is stale because it has been open 45 days with no activity. Remove stale label or comment or this will be closed in 10 days.'
close-pr-message: 'This PR was closed because it has been stalled for 10 days with no activity.'
exempt-pr-labels: 'needs_triage'
stale-pr-label: 'no_pr_activity'

View File

@@ -2,10 +2,12 @@
name: Feature branch deletion cleanup
env:
LC_ALL: "C.UTF-8" # prevent ERROR: Ansible could not initialize the preferred locale: unsupported locale setting
on: delete
on:
delete:
branches:
- feature_**
jobs:
branch_delete:
if: ${{ github.event.ref_type == 'branch' && startsWith(github.event.ref, 'feature_') }}
push:
runs-on: ubuntu-latest
timeout-minutes: 20
permissions:
@@ -20,4 +22,6 @@ jobs:
run: |
ansible localhost -c local, -m command -a "{{ ansible_python_interpreter + ' -m pip install boto3'}}"
ansible localhost -c local -m aws_s3 \
-a "bucket=awx-public-ci-files object=${GITHUB_REF##*/}/schema.json mode=delobj permission=public-read"
-a "bucket=awx-public-ci-files object=${GITHUB_REF##*/}/schema.json mode=delete permission=public-read"

View File

@@ -86,19 +86,13 @@ jobs:
-e push=yes \
-e awx_official=yes
- name: Log into registry ghcr.io
uses: docker/login-action@343f7c4344506bcbf9b4de18042ae17996df046d # v3.0.0
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Log in to GHCR
run: |
echo ${{ secrets.GITHUB_TOKEN }} | docker login ghcr.io -u ${{ github.actor }} --password-stdin
- name: Log into registry quay.io
uses: docker/login-action@343f7c4344506bcbf9b4de18042ae17996df046d # v3.0.0
with:
registry: quay.io
username: ${{ secrets.QUAY_USER }}
password: ${{ secrets.QUAY_TOKEN }}
- name: Log in to Quay
run: |
echo ${{ secrets.QUAY_TOKEN }} | docker login quay.io -u ${{ secrets.QUAY_USER }} --password-stdin
- name: tag awx-ee:latest with version input
run: |
@@ -106,13 +100,13 @@ jobs:
docker tag quay.io/ansible/awx-ee:latest ghcr.io/${{ github.repository_owner }}/awx-ee:${{ github.event.inputs.version }}
docker push ghcr.io/${{ github.repository_owner }}/awx-ee:${{ github.event.inputs.version }}
- name: Stage awx-operator image
- name: Build and stage awx-operator
working-directory: awx-operator
run: |
BUILD_ARGS="--build-arg DEFAULT_AWX_VERSION=${{ github.event.inputs.version}} \
--build-arg OPERATOR_VERSION=${{ github.event.inputs.operator_version }}" \
IMG=ghcr.io/${{ github.repository_owner }}/awx-operator:${{ github.event.inputs.operator_version }} \
make docker-buildx
BUILD_ARGS="--build-arg DEFAULT_AWX_VERSION=${{ github.event.inputs.version }} \
--build-arg OPERATOR_VERSION=${{ github.event.inputs.operator_version }}" \
IMAGE_TAG_BASE=ghcr.io/${{ github.repository_owner }}/awx-operator \
VERSION=${{ github.event.inputs.operator_version }} make docker-build docker-push
- name: Run test deployment with awx-operator
working-directory: awx-operator

3
.gitignore vendored
View File

@@ -169,6 +169,3 @@ awx/ui_next/build
# Docs build stuff
docs/docsite/build/
_readthedocs/
# Pyenv
.python-version

View File

@@ -10,7 +10,7 @@ KIND_BIN ?= $(shell which kind)
CHROMIUM_BIN=/tmp/chrome-linux/chrome
GIT_BRANCH ?= $(shell git rev-parse --abbrev-ref HEAD)
MANAGEMENT_COMMAND ?= awx-manage
VERSION ?= $(shell $(PYTHON) tools/scripts/scm_version.py 2> /dev/null)
VERSION ?= $(shell $(PYTHON) tools/scripts/scm_version.py)
# ansible-test requires semver compatable version, so we allow overrides to hack it
COLLECTION_VERSION ?= $(shell $(PYTHON) tools/scripts/scm_version.py | cut -d . -f 1-3)
@@ -538,8 +538,7 @@ docker-compose: awx/projects docker-compose-sources
ansible-galaxy install --ignore-certs -r tools/docker-compose/ansible/requirements.yml;
ansible-playbook -i tools/docker-compose/inventory tools/docker-compose/ansible/initialize_containers.yml \
-e enable_vault=$(VAULT) \
-e vault_tls=$(VAULT_TLS) \
-e enable_ldap=$(LDAP);
-e vault_tls=$(VAULT_TLS);
$(DOCKER_COMPOSE) -f tools/docker-compose/_sources/docker-compose.yml $(COMPOSE_OPTS) up $(COMPOSE_UP_OPTS) --remove-orphans
docker-compose-credential-plugins: awx/projects docker-compose-sources
@@ -586,35 +585,12 @@ docker-compose-build: Dockerfile.dev
--build-arg BUILDKIT_INLINE_CACHE=1 \
--cache-from=$(DEV_DOCKER_TAG_BASE)/awx_devel:$(COMPOSE_TAG) .
# ## Build awx_devel image for docker compose development environment for multiple architectures
# docker-compose-buildx: Dockerfile.dev
# DOCKER_BUILDKIT=1 docker build \
# -f Dockerfile.dev \
# -t $(DEVEL_IMAGE_NAME) \
# --build-arg BUILDKIT_INLINE_CACHE=1 \
# --cache-from=$(DEV_DOCKER_TAG_BASE)/awx_devel:$(COMPOSE_TAG) .
## Build awx_devel image for docker compose development environment for multiple architectures
# PLATFORMS defines the target platforms for the manager image be build to provide support to multiple
# architectures. (i.e. make docker-buildx IMG=myregistry/mypoperator:0.0.1). To use this option you need to:
# - able to use docker buildx . More info: https://docs.docker.com/build/buildx/
# - have enable BuildKit, More info: https://docs.docker.com/develop/develop-images/build_enhancements/
# - be able to push the image for your registry (i.e. if you do not inform a valid value via IMG=<myregistry/image:<tag>> than the export will fail)
# To properly provided solutions that supports more than one platform you should use this option.
PLATFORMS ?= linux/amd64,linux/arm64 # linux/ppc64le,linux/s390x
.PHONY: docker-compose-buildx
docker-compose-buildx: Dockerfile.dev ## Build and push docker image for the manager for cross-platform support
- docker buildx create --name project-v3-builder
docker buildx use project-v3-builder
- docker buildx build --push $(BUILD_ARGS) --platform=$(PLATFORMS) --tag $(DEVEL_IMAGE_NAME) -f Dockerfile.dev .
- docker buildx rm project-v3-builder
docker-clean:
-$(foreach container_id,$(shell docker ps -f name=tools_awx -aq && docker ps -f name=tools_receptor -aq),docker stop $(container_id); docker rm -f $(container_id);)
-$(foreach image_id,$(shell docker images --filter=reference='*/*/*awx_devel*' --filter=reference='*/*awx_devel*' --filter=reference='*awx_devel*' -aq),docker rmi --force $(image_id);)
docker-clean-volumes: docker-compose-clean docker-compose-container-group-clean
docker volume rm -f tools_awx_db tools_vault_1 tools_ldap_1 tools_grafana_storage tools_prometheus_storage $(docker volume ls --filter name=tools_redis_socket_ -q)
docker volume rm -f tools_awx_db tools_vault_1 tools_grafana_storage tools_prometheus_storage $(docker volume ls --filter name=tools_redis_socket_ -q)
docker-refresh: docker-clean docker-compose

View File

@@ -6,7 +6,7 @@ import copy
import json
import logging
import re
from collections import Counter, OrderedDict
from collections import OrderedDict
from datetime import timedelta
from uuid import uuid4
@@ -82,7 +82,6 @@ from awx.main.models import (
Project,
ProjectUpdate,
ProjectUpdateEvent,
ReceptorAddress,
RefreshToken,
Role,
Schedule,
@@ -637,7 +636,7 @@ class BaseSerializer(serializers.ModelSerializer, metaclass=BaseSerializerMetacl
exclusions = self.get_validation_exclusions(self.instance)
obj = self.instance or self.Meta.model()
for k, v in attrs.items():
if k not in exclusions and k != 'canonical_address_port':
if k not in exclusions:
setattr(obj, k, v)
obj.full_clean(exclude=exclusions)
# full_clean may modify values on the instance; copy those changes
@@ -5177,21 +5176,16 @@ class NotificationTemplateSerializer(BaseSerializer):
body = messages[event].get('body', {})
if body:
try:
sandbox.ImmutableSandboxedEnvironment(undefined=DescriptiveUndefined).from_string(body).render(JobNotificationMixin.context_stub())
# https://github.com/ansible/awx/issues/14410
# When rendering something such as "{{ job.id }}"
# the return type is not a dict, unlike "{{ job_metadata }}" which is a dict
# potential_body = json.loads(rendered_body)
# if not isinstance(potential_body, dict):
# error_list.append(
# _("Webhook body for '{}' should be a json dictionary. Found type '{}'.".format(event, type(potential_body).__name__))
# )
except Exception as exc:
error_list.append(_("Webhook body for '{}' is not valid. The following gave an error ({}).".format(event, exc)))
rendered_body = (
sandbox.ImmutableSandboxedEnvironment(undefined=DescriptiveUndefined).from_string(body).render(JobNotificationMixin.context_stub())
)
potential_body = json.loads(rendered_body)
if not isinstance(potential_body, dict):
error_list.append(
_("Webhook body for '{}' should be a json dictionary. Found type '{}'.".format(event, type(potential_body).__name__))
)
except json.JSONDecodeError as exc:
error_list.append(_("Webhook body for '{}' is not a valid json dictionary ({}).".format(event, exc)))
if error_list:
raise serializers.ValidationError(error_list)
@@ -5464,25 +5458,17 @@ class ScheduleSerializer(LaunchConfigurationBaseSerializer, SchedulePreviewSeria
class InstanceLinkSerializer(BaseSerializer):
class Meta:
model = InstanceLink
fields = ('id', 'related', 'source', 'target', 'target_full_address', 'link_state')
fields = ('id', 'url', 'related', 'source', 'target', 'link_state')
source = serializers.SlugRelatedField(slug_field="hostname", queryset=Instance.objects.all())
target = serializers.SerializerMethodField()
target_full_address = serializers.SerializerMethodField()
target = serializers.SlugRelatedField(slug_field="hostname", queryset=Instance.objects.all())
def get_related(self, obj):
res = super(InstanceLinkSerializer, self).get_related(obj)
res['source_instance'] = self.reverse('api:instance_detail', kwargs={'pk': obj.source.id})
res['target_address'] = self.reverse('api:receptor_address_detail', kwargs={'pk': obj.target.id})
res['target_instance'] = self.reverse('api:instance_detail', kwargs={'pk': obj.target.id})
return res
def get_target(self, obj):
return obj.target.instance.hostname
def get_target_full_address(self, obj):
return obj.target.get_full_address()
class InstanceNodeSerializer(BaseSerializer):
class Meta:
@@ -5490,29 +5476,6 @@ class InstanceNodeSerializer(BaseSerializer):
fields = ('id', 'hostname', 'node_type', 'node_state', 'enabled')
class ReceptorAddressSerializer(BaseSerializer):
full_address = serializers.SerializerMethodField()
class Meta:
model = ReceptorAddress
fields = (
'id',
'url',
'address',
'port',
'protocol',
'websocket_path',
'is_internal',
'canonical',
'instance',
'peers_from_control_nodes',
'full_address',
)
def get_full_address(self, obj):
return obj.get_full_address()
class InstanceSerializer(BaseSerializer):
show_capabilities = ['edit']
@@ -5521,17 +5484,11 @@ class InstanceSerializer(BaseSerializer):
jobs_running = serializers.IntegerField(help_text=_('Count of jobs in the running or waiting state that are targeted for this instance'), read_only=True)
jobs_total = serializers.IntegerField(help_text=_('Count of all jobs that target this instance'), read_only=True)
health_check_pending = serializers.SerializerMethodField()
peers = serializers.PrimaryKeyRelatedField(
help_text=_('Primary keys of receptor addresses to peer to.'), many=True, required=False, queryset=ReceptorAddress.objects.all()
)
reverse_peers = serializers.SerializerMethodField()
listener_port = serializers.IntegerField(source='canonical_address_port', required=False, allow_null=True)
peers_from_control_nodes = serializers.BooleanField(source='canonical_address_peers_from_control_nodes', required=False)
protocol = serializers.SerializerMethodField()
peers = serializers.SlugRelatedField(many=True, required=False, slug_field="hostname", queryset=Instance.objects.all())
class Meta:
model = Instance
read_only_fields = ('ip_address', 'uuid', 'version', 'managed', 'reverse_peers')
read_only_fields = ('ip_address', 'uuid', 'version')
fields = (
'id',
'hostname',
@@ -5562,13 +5519,10 @@ class InstanceSerializer(BaseSerializer):
'managed_by_policy',
'node_type',
'node_state',
'managed',
'ip_address',
'peers',
'reverse_peers',
'listener_port',
'peers',
'peers_from_control_nodes',
'protocol',
)
extra_kwargs = {
'node_type': {'initial': Instance.Types.EXECUTION, 'default': Instance.Types.EXECUTION},
@@ -5590,54 +5544,16 @@ class InstanceSerializer(BaseSerializer):
def get_related(self, obj):
res = super(InstanceSerializer, self).get_related(obj)
res['receptor_addresses'] = self.reverse('api:instance_receptor_addresses_list', kwargs={'pk': obj.pk})
res['jobs'] = self.reverse('api:instance_unified_jobs_list', kwargs={'pk': obj.pk})
res['peers'] = self.reverse('api:instance_peers_list', kwargs={"pk": obj.pk})
res['instance_groups'] = self.reverse('api:instance_instance_groups_list', kwargs={'pk': obj.pk})
if obj.node_type in [Instance.Types.EXECUTION, Instance.Types.HOP]:
res['install_bundle'] = self.reverse('api:instance_install_bundle', kwargs={'pk': obj.pk})
res['peers'] = self.reverse('api:instance_peers_list', kwargs={"pk": obj.pk})
if self.context['request'].user.is_superuser or self.context['request'].user.is_system_auditor:
if obj.node_type == 'execution':
res['health_check'] = self.reverse('api:instance_health_check', kwargs={'pk': obj.pk})
return res
def create_or_update(self, validated_data, obj=None, create=True):
# create a managed receptor address if listener port is defined
port = validated_data.pop('listener_port', -1)
peers_from_control_nodes = validated_data.pop('peers_from_control_nodes', -1)
# delete the receptor address if the port is explicitly set to None
if obj and port == None:
obj.receptor_addresses.filter(address=obj.hostname).delete()
if create:
instance = super(InstanceSerializer, self).create(validated_data)
else:
instance = super(InstanceSerializer, self).update(obj, validated_data)
instance.refresh_from_db() # instance canonical address lookup is deferred, so needs to be reloaded
# only create or update if port is defined in validated_data or already exists in the
# canonical address
# this prevents creating a receptor address if peers_from_control_nodes is in
# validated_data but a port is not set
if (port != None and port != -1) or instance.canonical_address_port:
kwargs = {}
if port != -1:
kwargs['port'] = port
if peers_from_control_nodes != -1:
kwargs['peers_from_control_nodes'] = peers_from_control_nodes
if kwargs:
kwargs['canonical'] = True
instance.receptor_addresses.update_or_create(address=instance.hostname, defaults=kwargs)
return instance
def create(self, validated_data):
return self.create_or_update(validated_data, create=True)
def update(self, obj, validated_data):
return self.create_or_update(validated_data, obj, create=False)
def get_summary_fields(self, obj):
summary = super().get_summary_fields(obj)
@@ -5647,16 +5563,6 @@ class InstanceSerializer(BaseSerializer):
return summary
def get_reverse_peers(self, obj):
return Instance.objects.prefetch_related('peers').filter(peers__in=obj.receptor_addresses.all()).values_list('id', flat=True)
def get_protocol(self, obj):
# note: don't create a different query for receptor addresses, as this is prefetched on the View for optimization
for addr in obj.receptor_addresses.all():
if addr.canonical:
return addr.protocol
return ""
def get_consumed_capacity(self, obj):
return obj.consumed_capacity
@@ -5670,20 +5576,47 @@ class InstanceSerializer(BaseSerializer):
return obj.health_check_pending
def validate(self, attrs):
# Oddly, using 'source' on a DRF field populates attrs with the source name, so we should rename it back
if 'canonical_address_port' in attrs:
attrs['listener_port'] = attrs.pop('canonical_address_port')
if 'canonical_address_peers_from_control_nodes' in attrs:
attrs['peers_from_control_nodes'] = attrs.pop('canonical_address_peers_from_control_nodes')
def get_field_from_model_or_attrs(fd):
return attrs.get(fd, self.instance and getattr(self.instance, fd) or None)
def check_peers_changed():
'''
return True if
- 'peers' in attrs
- instance peers matches peers in attrs
'''
return self.instance and 'peers' in attrs and set(self.instance.peers.all()) != set(attrs['peers'])
if not self.instance and not settings.IS_K8S:
raise serializers.ValidationError(_("Can only create instances on Kubernetes or OpenShift."))
# cannot enable peers_from_control_nodes if listener_port is not set
if attrs.get('peers_from_control_nodes'):
port = attrs.get('listener_port', -1) # -1 denotes missing, None denotes explicit null
if (port is None) or (port == -1 and self.instance and self.instance.canonical_address is None):
raise serializers.ValidationError(_("Cannot enable peers_from_control_nodes if listener_port is not set."))
node_type = get_field_from_model_or_attrs("node_type")
peers_from_control_nodes = get_field_from_model_or_attrs("peers_from_control_nodes")
listener_port = get_field_from_model_or_attrs("listener_port")
peers = attrs.get('peers', [])
if peers_from_control_nodes and node_type not in (Instance.Types.EXECUTION, Instance.Types.HOP):
raise serializers.ValidationError(_("peers_from_control_nodes can only be enabled for execution or hop nodes."))
if node_type in [Instance.Types.CONTROL, Instance.Types.HYBRID]:
if check_peers_changed():
raise serializers.ValidationError(
_("Setting peers manually for control nodes is not allowed. Enable peers_from_control_nodes on the hop and execution nodes instead.")
)
if not listener_port and peers_from_control_nodes:
raise serializers.ValidationError(_("Field listener_port must be a valid integer when peers_from_control_nodes is enabled."))
if not listener_port and self.instance and self.instance.peers_from.exists():
raise serializers.ValidationError(_("Field listener_port must be a valid integer when other nodes peer to it."))
for peer in peers:
if peer.listener_port is None:
raise serializers.ValidationError(_("Field listener_port must be set on peer ") + peer.hostname + ".")
if not settings.IS_K8S:
if check_peers_changed():
raise serializers.ValidationError(_("Cannot change peers."))
return super().validate(attrs)
@@ -5703,8 +5636,8 @@ class InstanceSerializer(BaseSerializer):
raise serializers.ValidationError(_("Can only change the state on Kubernetes or OpenShift."))
if value != Instance.States.DEPROVISIONING:
raise serializers.ValidationError(_("Can only change instances to the 'deprovisioning' state."))
if self.instance.managed:
raise serializers.ValidationError(_("Cannot deprovision managed nodes."))
if self.instance.node_type not in (Instance.Types.EXECUTION, Instance.Types.HOP):
raise serializers.ValidationError(_("Can only deprovision execution or hop nodes."))
else:
if value and value != Instance.States.INSTALLED:
raise serializers.ValidationError(_("Can only create instances in the 'installed' state."))
@@ -5723,48 +5656,18 @@ class InstanceSerializer(BaseSerializer):
def validate_listener_port(self, value):
"""
Cannot change listener port, unless going from none to integer, and vice versa
If instance is managed, cannot change listener port at all
"""
if self.instance:
canonical_address_port = self.instance.canonical_address_port
if value and canonical_address_port and canonical_address_port != value:
raise serializers.ValidationError(_("Cannot change listener port."))
if self.instance.managed and value != canonical_address_port:
raise serializers.ValidationError(_("Cannot change listener port for managed nodes."))
return value
def validate_peers(self, value):
# cannot peer to an instance more than once
peers_instances = Counter(p.instance_id for p in value)
if any(count > 1 for count in peers_instances.values()):
raise serializers.ValidationError(_("Cannot peer to the same instance more than once."))
if self.instance:
instance_addresses = set(self.instance.receptor_addresses.all())
setting_peers = set(value)
peers_changed = set(self.instance.peers.all()) != setting_peers
if not settings.IS_K8S and peers_changed:
raise serializers.ValidationError(_("Cannot change peers."))
if self.instance.managed and peers_changed:
raise serializers.ValidationError(_("Setting peers manually for managed nodes is not allowed."))
# cannot peer to self
if instance_addresses & setting_peers:
raise serializers.ValidationError(_("Instance cannot peer to its own address."))
# cannot peer to an instance that is already peered to this instance
if instance_addresses:
for p in setting_peers:
if set(p.instance.peers.all()) & instance_addresses:
raise serializers.ValidationError(_(f"Instance {p.instance.hostname} is already peered to this instance."))
if value and self.instance and self.instance.listener_port and self.instance.listener_port != value:
raise serializers.ValidationError(_("Cannot change listener port."))
return value
def validate_peers_from_control_nodes(self, value):
if self.instance and self.instance.managed and self.instance.canonical_address_peers_from_control_nodes != value:
raise serializers.ValidationError(_("Cannot change peers_from_control_nodes for managed nodes."))
"""
Can only enable for K8S based deployments
"""
if value and not settings.IS_K8S:
raise serializers.ValidationError(_("Can only be enabled on Kubernetes or Openshift."))
return value

View File

@@ -17,18 +17,19 @@ custom_worksign_public_keyfile: receptor/work_public_key.pem
custom_tls_certfile: receptor/tls/receptor.crt
custom_tls_keyfile: receptor/tls/receptor.key
custom_ca_certfile: receptor/tls/ca/mesh-CA.crt
{% if listener_port %}
receptor_protocol: {{ listener_protocol }}
receptor_protocol: 'tcp'
{% if instance.listener_port %}
receptor_listener: true
receptor_port: {{ listener_port }}
receptor_port: {{ instance.listener_port }}
{% else %}
receptor_listener: false
{% endif %}
{% if peers %}
receptor_peers:
{% for peer in peers %}
- address: {{ peer.address }}
protocol: {{ peer.protocol }}
- host: {{ peer.host }}
port: {{ peer.port }}
protocol: tcp
{% endfor %}
{% endif %}
{% verbatim %}

View File

@@ -1,4 +1,4 @@
---
collections:
- name: ansible.receptor
version: 2.0.3
version: 2.0.2

View File

@@ -10,7 +10,6 @@ from awx.api.views import (
InstanceInstanceGroupsList,
InstanceHealthCheck,
InstancePeersList,
InstanceReceptorAddressesList,
)
from awx.api.views.instance_install_bundle import InstanceInstallBundle
@@ -22,7 +21,6 @@ urls = [
re_path(r'^(?P<pk>[0-9]+)/instance_groups/$', InstanceInstanceGroupsList.as_view(), name='instance_instance_groups_list'),
re_path(r'^(?P<pk>[0-9]+)/health_check/$', InstanceHealthCheck.as_view(), name='instance_health_check'),
re_path(r'^(?P<pk>[0-9]+)/peers/$', InstancePeersList.as_view(), name='instance_peers_list'),
re_path(r'^(?P<pk>[0-9]+)/receptor_addresses/$', InstanceReceptorAddressesList.as_view(), name='instance_receptor_addresses_list'),
re_path(r'^(?P<pk>[0-9]+)/install_bundle/$', InstanceInstallBundle.as_view(), name='instance_install_bundle'),
]

View File

@@ -1,17 +0,0 @@
# Copyright (c) 2017 Ansible, Inc.
# All Rights Reserved.
from django.urls import re_path
from awx.api.views import (
ReceptorAddressesList,
ReceptorAddressDetail,
)
urls = [
re_path(r'^$', ReceptorAddressesList.as_view(), name='receptor_addresses_list'),
re_path(r'^(?P<pk>[0-9]+)/$', ReceptorAddressDetail.as_view(), name='receptor_address_detail'),
]
__all__ = ['urls']

View File

@@ -85,7 +85,6 @@ from .oauth2_root import urls as oauth2_root_urls
from .workflow_approval_template import urls as workflow_approval_template_urls
from .workflow_approval import urls as workflow_approval_urls
from .analytics import urls as analytics_urls
from .receptor_address import urls as receptor_address_urls
v2_urls = [
re_path(r'^$', ApiV2RootView.as_view(), name='api_v2_root_view'),
@@ -156,7 +155,6 @@ v2_urls = [
re_path(r'^bulk/host_create/$', BulkHostCreateView.as_view(), name='bulk_host_create'),
re_path(r'^bulk/host_delete/$', BulkHostDeleteView.as_view(), name='bulk_host_delete'),
re_path(r'^bulk/job_launch/$', BulkJobLaunchView.as_view(), name='bulk_job_launch'),
re_path(r'^receptor_addresses/', include(receptor_address_urls)),
]

View File

@@ -337,20 +337,12 @@ class InstanceList(ListCreateAPIView):
search_fields = ('hostname',)
ordering = ('id',)
def get_queryset(self):
qs = super().get_queryset().prefetch_related('receptor_addresses')
return qs
class InstanceDetail(RetrieveUpdateAPIView):
name = _("Instance Detail")
model = models.Instance
serializer_class = serializers.InstanceSerializer
def get_queryset(self):
qs = super().get_queryset().prefetch_related('receptor_addresses')
return qs
def update_raw_data(self, data):
# these fields are only valid on creation of an instance, so they unwanted on detail view
data.pop('node_type', None)
@@ -383,37 +375,13 @@ class InstanceUnifiedJobsList(SubListAPIView):
class InstancePeersList(SubListAPIView):
name = _("Peers")
model = models.ReceptorAddress
serializer_class = serializers.ReceptorAddressSerializer
name = _("Instance Peers")
parent_model = models.Instance
model = models.Instance
serializer_class = serializers.InstanceSerializer
parent_access = 'read'
search_fields = {'hostname'}
relationship = 'peers'
search_fields = ('address',)
class InstanceReceptorAddressesList(SubListAPIView):
name = _("Receptor Addresses")
model = models.ReceptorAddress
parent_key = 'instance'
parent_model = models.Instance
serializer_class = serializers.ReceptorAddressSerializer
search_fields = ('address',)
class ReceptorAddressesList(ListAPIView):
name = _("Receptor Addresses")
model = models.ReceptorAddress
serializer_class = serializers.ReceptorAddressSerializer
search_fields = ('address',)
class ReceptorAddressDetail(RetrieveAPIView):
name = _("Receptor Address Detail")
model = models.ReceptorAddress
serializer_class = serializers.ReceptorAddressSerializer
parent_model = models.Instance
relationship = 'receptor_addresses'
class InstanceInstanceGroupsList(InstanceGroupMembershipMixin, SubListCreateAttachDetachAPIView):

View File

@@ -124,19 +124,10 @@ def generate_inventory_yml(instance_obj):
def generate_group_vars_all_yml(instance_obj):
# get peers
peers = []
for addr in instance_obj.peers.select_related('instance'):
peers.append(dict(address=addr.get_full_address(), protocol=addr.protocol))
context = dict(instance=instance_obj, peers=peers)
canonical_addr = instance_obj.canonical_address
if canonical_addr:
context['listener_port'] = canonical_addr.port
protocol = canonical_addr.protocol if canonical_addr.protocol != 'wss' else 'ws'
context['listener_protocol'] = protocol
all_yaml = render_to_string("instance_install_bundle/group_vars/all.yml", context=context)
for instance in instance_obj.peers.all():
peers.append(dict(host=instance.hostname, port=instance.listener_port))
all_yaml = render_to_string("instance_install_bundle/group_vars/all.yml", context=dict(instance=instance_obj, peers=peers))
# convert consecutive newlines with a single newline
return re.sub(r'\n+', '\n', all_yaml)

View File

@@ -17,7 +17,7 @@ class MeshVisualizer(APIView):
def get(self, request, format=None):
data = {
'nodes': InstanceNodeSerializer(Instance.objects.all(), many=True).data,
'links': InstanceLinkSerializer(InstanceLink.objects.select_related('target__instance', 'source'), many=True).data,
'links': InstanceLinkSerializer(InstanceLink.objects.select_related('target', 'source'), many=True).data,
}
return Response(data)

View File

@@ -84,7 +84,6 @@ class ApiVersionRootView(APIView):
data['ping'] = reverse('api:api_v2_ping_view', request=request)
data['instances'] = reverse('api:instance_list', request=request)
data['instance_groups'] = reverse('api:instance_group_list', request=request)
data['receptor_addresses'] = reverse('api:receptor_addresses_list', request=request)
data['config'] = reverse('api:api_v2_config_view', request=request)
data['settings'] = reverse('api:setting_category_list', request=request)
data['me'] = reverse('api:user_me_list', request=request)

View File

@@ -57,7 +57,6 @@ from awx.main.models import (
Project,
ProjectUpdate,
ProjectUpdateEvent,
ReceptorAddress,
Role,
Schedule,
SystemJob,
@@ -2431,29 +2430,6 @@ class InventoryUpdateEventAccess(BaseAccess):
return False
class ReceptorAddressAccess(BaseAccess):
"""
I can see receptor address records whenever I can access the instance
"""
model = ReceptorAddress
def filtered_queryset(self):
return self.model.objects.filter(Q(instance__in=Instance.accessible_pk_qs(self.user, 'read_role')))
@check_superuser
def can_add(self, data):
return False
@check_superuser
def can_change(self, obj, data):
return False
@check_superuser
def can_delete(self, obj):
return False
class SystemJobEventAccess(BaseAccess):
"""
I can only see manage System Jobs events if I'm a super user

View File

@@ -2,7 +2,7 @@
import logging
# AWX
from awx.main.analytics.subsystem_metrics import DispatcherMetrics, CallbackReceiverMetrics
from awx.main.analytics.subsystem_metrics import Metrics
from awx.main.dispatch.publish import task
from awx.main.dispatch import get_task_queuename
@@ -11,5 +11,4 @@ logger = logging.getLogger('awx.main.scheduler')
@task(queue=get_task_queuename)
def send_subsystem_metrics():
DispatcherMetrics().send_metrics()
CallbackReceiverMetrics().send_metrics()
Metrics().send_metrics()

View File

@@ -1,15 +1,10 @@
import itertools
import redis
import json
import time
import logging
import prometheus_client
from prometheus_client.core import GaugeMetricFamily, HistogramMetricFamily
from prometheus_client.registry import CollectorRegistry
from django.conf import settings
from django.http import HttpRequest
from rest_framework.request import Request
from django.apps import apps
from awx.main.consumers import emit_channel_notification
from awx.main.utils import is_testing
@@ -18,30 +13,6 @@ root_key = settings.SUBSYSTEM_METRICS_REDIS_KEY_PREFIX
logger = logging.getLogger('awx.main.analytics')
class MetricsNamespace:
def __init__(self, namespace):
self._namespace = namespace
class MetricsServerSettings(MetricsNamespace):
def port(self):
return settings.METRICS_SUBSYSTEM_CONFIG['server'][self._namespace]['port']
class MetricsServer(MetricsServerSettings):
def __init__(self, namespace, registry):
MetricsNamespace.__init__(self, namespace)
self._registry = registry
def start(self):
try:
# TODO: addr for ipv6 ?
prometheus_client.start_http_server(self.port(), addr='localhost', registry=self._registry)
except Exception:
logger.error(f"MetricsServer failed to start for service '{self._namespace}.")
raise
class BaseM:
def __init__(self, field, help_text):
self.field = field
@@ -177,40 +148,76 @@ class HistogramM(BaseM):
return output_text
class Metrics(MetricsNamespace):
# metric name, help_text
METRICSLIST = []
_METRICSLIST = [
FloatM('subsystem_metrics_pipe_execute_seconds', 'Time spent saving metrics to redis'),
IntM('subsystem_metrics_pipe_execute_calls', 'Number of calls to pipe_execute'),
FloatM('subsystem_metrics_send_metrics_seconds', 'Time spent sending metrics to other nodes'),
]
def __init__(self, namespace, auto_pipe_execute=False, instance_name=None, metrics_have_changed=True, **kwargs):
MetricsNamespace.__init__(self, namespace)
class Metrics:
def __init__(self, auto_pipe_execute=False, instance_name=None):
self.pipe = redis.Redis.from_url(settings.BROKER_URL).pipeline()
self.conn = redis.Redis.from_url(settings.BROKER_URL)
self.last_pipe_execute = time.time()
# track if metrics have been modified since last saved to redis
# start with True so that we get an initial save to redis
self.metrics_have_changed = metrics_have_changed
self.metrics_have_changed = True
self.pipe_execute_interval = settings.SUBSYSTEM_METRICS_INTERVAL_SAVE_TO_REDIS
self.send_metrics_interval = settings.SUBSYSTEM_METRICS_INTERVAL_SEND_METRICS
# auto pipe execute will commit transaction of metric data to redis
# at a regular interval (pipe_execute_interval). If set to False,
# the calling function should call .pipe_execute() explicitly
self.auto_pipe_execute = auto_pipe_execute
Instance = apps.get_model('main', 'Instance')
if instance_name:
self.instance_name = instance_name
elif is_testing():
self.instance_name = "awx_testing"
else:
self.instance_name = settings.CLUSTER_HOST_ID # Same as Instance.objects.my_hostname() BUT we do not need to import Instance
self.instance_name = Instance.objects.my_hostname()
# metric name, help_text
METRICSLIST = [
SetIntM('callback_receiver_events_queue_size_redis', 'Current number of events in redis queue'),
IntM('callback_receiver_events_popped_redis', 'Number of events popped from redis'),
IntM('callback_receiver_events_in_memory', 'Current number of events in memory (in transfer from redis to db)'),
IntM('callback_receiver_batch_events_errors', 'Number of times batch insertion failed'),
FloatM('callback_receiver_events_insert_db_seconds', 'Total time spent saving events to database'),
IntM('callback_receiver_events_insert_db', 'Number of events batch inserted into database'),
IntM('callback_receiver_events_broadcast', 'Number of events broadcast to other control plane nodes'),
HistogramM(
'callback_receiver_batch_events_insert_db', 'Number of events batch inserted into database', settings.SUBSYSTEM_METRICS_BATCH_INSERT_BUCKETS
),
SetFloatM('callback_receiver_event_processing_avg_seconds', 'Average processing time per event per callback receiver batch'),
FloatM('subsystem_metrics_pipe_execute_seconds', 'Time spent saving metrics to redis'),
IntM('subsystem_metrics_pipe_execute_calls', 'Number of calls to pipe_execute'),
FloatM('subsystem_metrics_send_metrics_seconds', 'Time spent sending metrics to other nodes'),
SetFloatM('task_manager_get_tasks_seconds', 'Time spent in loading tasks from db'),
SetFloatM('task_manager_start_task_seconds', 'Time spent starting task'),
SetFloatM('task_manager_process_running_tasks_seconds', 'Time spent processing running tasks'),
SetFloatM('task_manager_process_pending_tasks_seconds', 'Time spent processing pending tasks'),
SetFloatM('task_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('task_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('task_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetIntM('task_manager_tasks_started', 'Number of tasks started'),
SetIntM('task_manager_running_processed', 'Number of running tasks processed'),
SetIntM('task_manager_pending_processed', 'Number of pending tasks processed'),
SetIntM('task_manager_tasks_blocked', 'Number of tasks blocked from running'),
SetFloatM('task_manager_commit_seconds', 'Time spent in db transaction, including on_commit calls'),
SetFloatM('dependency_manager_get_tasks_seconds', 'Time spent loading pending tasks from db'),
SetFloatM('dependency_manager_generate_dependencies_seconds', 'Time spent generating dependencies for pending tasks'),
SetFloatM('dependency_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('dependency_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('dependency_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetIntM('dependency_manager_pending_processed', 'Number of pending tasks processed'),
SetFloatM('workflow_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('workflow_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('workflow_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetFloatM('workflow_manager_spawn_workflow_graph_jobs_seconds', 'Time spent spawning workflow tasks'),
SetFloatM('workflow_manager_get_tasks_seconds', 'Time spent loading workflow tasks from db'),
# dispatcher subsystem metrics
SetIntM('dispatcher_pool_scale_up_events', 'Number of times local dispatcher scaled up a worker since startup'),
SetIntM('dispatcher_pool_active_task_count', 'Number of active tasks in the worker pool when last task was submitted'),
SetIntM('dispatcher_pool_max_worker_count', 'Highest number of workers in worker pool in last collection interval, about 20s'),
SetFloatM('dispatcher_availability', 'Fraction of time (in last collection interval) dispatcher was able to receive messages'),
]
# turn metric list into dictionary with the metric name as a key
self.METRICS = {}
for m in itertools.chain(self.METRICSLIST, self._METRICSLIST):
for m in METRICSLIST:
self.METRICS[m.field] = m
# track last time metrics were sent to other nodes
@@ -223,7 +230,7 @@ class Metrics(MetricsNamespace):
m.reset_value(self.conn)
self.metrics_have_changed = True
self.conn.delete(root_key + "_lock")
for m in self.conn.scan_iter(root_key + '-' + self._namespace + '_instance_*'):
for m in self.conn.scan_iter(root_key + '_instance_*'):
self.conn.delete(m)
def inc(self, field, value):
@@ -290,7 +297,7 @@ class Metrics(MetricsNamespace):
def send_metrics(self):
# more than one thread could be calling this at the same time, so should
# acquire redis lock before sending metrics
lock = self.conn.lock(root_key + '-' + self._namespace + '_lock')
lock = self.conn.lock(root_key + '_lock')
if not lock.acquire(blocking=False):
return
try:
@@ -300,10 +307,9 @@ class Metrics(MetricsNamespace):
payload = {
'instance': self.instance_name,
'metrics': serialized_metrics,
'metrics_namespace': self._namespace,
}
# store the serialized data locally as well, so that load_other_metrics will read it
self.conn.set(root_key + '-' + self._namespace + '_instance_' + self.instance_name, serialized_metrics)
self.conn.set(root_key + '_instance_' + self.instance_name, serialized_metrics)
emit_channel_notification("metrics", payload)
self.previous_send_metrics.set(current_time)
@@ -325,14 +331,14 @@ class Metrics(MetricsNamespace):
instances_filter = request.query_params.getlist("node")
# get a sorted list of instance names
instance_names = [self.instance_name]
for m in self.conn.scan_iter(root_key + '-' + self._namespace + '_instance_*'):
for m in self.conn.scan_iter(root_key + '_instance_*'):
instance_names.append(m.decode('UTF-8').split('_instance_')[1])
instance_names.sort()
# load data, including data from the this local instance
instance_data = {}
for instance in instance_names:
if len(instances_filter) == 0 or instance in instances_filter:
instance_data_from_redis = self.conn.get(root_key + '-' + self._namespace + '_instance_' + instance)
instance_data_from_redis = self.conn.get(root_key + '_instance_' + instance)
# data from other instances may not be available. That is OK.
if instance_data_from_redis:
instance_data[instance] = json.loads(instance_data_from_redis.decode('UTF-8'))
@@ -351,120 +357,6 @@ class Metrics(MetricsNamespace):
return output_text
class DispatcherMetrics(Metrics):
METRICSLIST = [
SetFloatM('task_manager_get_tasks_seconds', 'Time spent in loading tasks from db'),
SetFloatM('task_manager_start_task_seconds', 'Time spent starting task'),
SetFloatM('task_manager_process_running_tasks_seconds', 'Time spent processing running tasks'),
SetFloatM('task_manager_process_pending_tasks_seconds', 'Time spent processing pending tasks'),
SetFloatM('task_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('task_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('task_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetIntM('task_manager_tasks_started', 'Number of tasks started'),
SetIntM('task_manager_running_processed', 'Number of running tasks processed'),
SetIntM('task_manager_pending_processed', 'Number of pending tasks processed'),
SetIntM('task_manager_tasks_blocked', 'Number of tasks blocked from running'),
SetFloatM('task_manager_commit_seconds', 'Time spent in db transaction, including on_commit calls'),
SetFloatM('dependency_manager_get_tasks_seconds', 'Time spent loading pending tasks from db'),
SetFloatM('dependency_manager_generate_dependencies_seconds', 'Time spent generating dependencies for pending tasks'),
SetFloatM('dependency_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('dependency_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('dependency_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetIntM('dependency_manager_pending_processed', 'Number of pending tasks processed'),
SetFloatM('workflow_manager__schedule_seconds', 'Time spent in running the entire _schedule'),
IntM('workflow_manager__schedule_calls', 'Number of calls to _schedule, after lock is acquired'),
SetFloatM('workflow_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
SetFloatM('workflow_manager_spawn_workflow_graph_jobs_seconds', 'Time spent spawning workflow tasks'),
SetFloatM('workflow_manager_get_tasks_seconds', 'Time spent loading workflow tasks from db'),
# dispatcher subsystem metrics
SetIntM('dispatcher_pool_scale_up_events', 'Number of times local dispatcher scaled up a worker since startup'),
SetIntM('dispatcher_pool_active_task_count', 'Number of active tasks in the worker pool when last task was submitted'),
SetIntM('dispatcher_pool_max_worker_count', 'Highest number of workers in worker pool in last collection interval, about 20s'),
SetFloatM('dispatcher_availability', 'Fraction of time (in last collection interval) dispatcher was able to receive messages'),
]
def __init__(self, *args, **kwargs):
super().__init__(settings.METRICS_SERVICE_DISPATCHER, *args, **kwargs)
class CallbackReceiverMetrics(Metrics):
METRICSLIST = [
SetIntM('callback_receiver_events_queue_size_redis', 'Current number of events in redis queue'),
IntM('callback_receiver_events_popped_redis', 'Number of events popped from redis'),
IntM('callback_receiver_events_in_memory', 'Current number of events in memory (in transfer from redis to db)'),
IntM('callback_receiver_batch_events_errors', 'Number of times batch insertion failed'),
FloatM('callback_receiver_events_insert_db_seconds', 'Total time spent saving events to database'),
IntM('callback_receiver_events_insert_db', 'Number of events batch inserted into database'),
IntM('callback_receiver_events_broadcast', 'Number of events broadcast to other control plane nodes'),
HistogramM(
'callback_receiver_batch_events_insert_db', 'Number of events batch inserted into database', settings.SUBSYSTEM_METRICS_BATCH_INSERT_BUCKETS
),
SetFloatM('callback_receiver_event_processing_avg_seconds', 'Average processing time per event per callback receiver batch'),
]
def __init__(self, *args, **kwargs):
super().__init__(settings.METRICS_SERVICE_CALLBACK_RECEIVER, *args, **kwargs)
def metrics(request):
output_text = ''
for m in [DispatcherMetrics(), CallbackReceiverMetrics()]:
output_text += m.generate_metrics(request)
return output_text
class CustomToPrometheusMetricsCollector(prometheus_client.registry.Collector):
"""
Takes the metric data from redis -> our custom metric fields -> prometheus
library metric fields.
The plan is to get rid of the use of redis, our custom metric fields, and
to switch fully to the prometheus library. At that point, this translation
code will be deleted.
"""
def __init__(self, metrics_obj, *args, **kwargs):
super().__init__(*args, **kwargs)
self._metrics = metrics_obj
def collect(self):
my_hostname = settings.CLUSTER_HOST_ID
instance_data = self._metrics.load_other_metrics(Request(HttpRequest()))
if not instance_data:
logger.debug(f"No metric data not found in redis for metric namespace '{self._metrics._namespace}'")
return None
host_metrics = instance_data.get(my_hostname)
for _, metric in self._metrics.METRICS.items():
entry = host_metrics.get(metric.field)
if not entry:
logger.debug(f"{self._metrics._namespace} metric '{metric.field}' not found in redis data payload {json.dumps(instance_data, indent=2)}")
continue
if isinstance(metric, HistogramM):
buckets = list(zip(metric.buckets, entry['counts']))
buckets = [[str(i[0]), str(i[1])] for i in buckets]
yield HistogramMetricFamily(metric.field, metric.help_text, buckets=buckets, sum_value=entry['sum'])
else:
yield GaugeMetricFamily(metric.field, metric.help_text, value=entry)
class CallbackReceiverMetricsServer(MetricsServer):
def __init__(self):
registry = CollectorRegistry(auto_describe=True)
registry.register(CustomToPrometheusMetricsCollector(DispatcherMetrics(metrics_have_changed=False)))
super().__init__(settings.METRICS_SERVICE_CALLBACK_RECEIVER, registry)
class DispatcherMetricsServer(MetricsServer):
def __init__(self):
registry = CollectorRegistry(auto_describe=True)
registry.register(CustomToPrometheusMetricsCollector(CallbackReceiverMetrics(metrics_have_changed=False)))
super().__init__(settings.METRICS_SERVICE_DISPATCHER, registry)
class WebsocketsMetricsServer(MetricsServer):
def __init__(self):
registry = CollectorRegistry(auto_describe=True)
# registry.register()
super().__init__(settings.METRICS_SERVICE_WEBSOCKETS, registry)
m = Metrics()
return m.generate_metrics(request)

View File

@@ -106,7 +106,7 @@ class RelayConsumer(AsyncJsonWebsocketConsumer):
if group == "metrics":
message = json.loads(message['text'])
conn = redis.Redis.from_url(settings.BROKER_URL)
conn.set(settings.SUBSYSTEM_METRICS_REDIS_KEY_PREFIX + "-" + message['metrics_namespace'] + "_instance_" + message['instance'], message['metrics'])
conn.set(settings.SUBSYSTEM_METRICS_REDIS_KEY_PREFIX + "_instance_" + message['instance'], message['metrics'])
else:
await self.channel_layer.group_send(group, message)

View File

@@ -105,11 +105,7 @@ def create_listener_connection():
for k, v in settings.LISTENER_DATABASES.get('default', {}).get('OPTIONS', {}).items():
conf['OPTIONS'][k] = v
# Allow password-less authentication
if 'PASSWORD' in conf:
conf['OPTIONS']['password'] = conf.pop('PASSWORD')
connection_data = f"dbname={conf['NAME']} host={conf['HOST']} user={conf['USER']} port={conf['PORT']}"
connection_data = f"dbname={conf['NAME']} host={conf['HOST']} user={conf['USER']} password={conf['PASSWORD']} port={conf['PORT']}"
return psycopg.connect(connection_data, autocommit=True, **conf['OPTIONS'])

View File

@@ -162,13 +162,13 @@ class AWXConsumerRedis(AWXConsumerBase):
class AWXConsumerPG(AWXConsumerBase):
def __init__(self, *args, schedule=None, **kwargs):
super().__init__(*args, **kwargs)
self.pg_max_wait = getattr(settings, 'DISPATCHER_DB_DOWNTOWN_TOLLERANCE', settings.DISPATCHER_DB_DOWNTIME_TOLERANCE)
self.pg_max_wait = settings.DISPATCHER_DB_DOWNTIME_TOLERANCE
# if no successful loops have ran since startup, then we should fail right away
self.pg_is_down = True # set so that we fail if we get database errors on startup
init_time = time.time()
self.pg_down_time = init_time - self.pg_max_wait # allow no grace period
self.last_cleanup = init_time
self.subsystem_metrics = s_metrics.DispatcherMetrics(auto_pipe_execute=False)
self.subsystem_metrics = s_metrics.Metrics(auto_pipe_execute=False)
self.last_metrics_gather = init_time
self.listen_cumulative_time = 0.0
if schedule:

View File

@@ -72,7 +72,7 @@ class CallbackBrokerWorker(BaseWorker):
def __init__(self):
self.buff = {}
self.redis = redis.Redis.from_url(settings.BROKER_URL)
self.subsystem_metrics = s_metrics.CallbackReceiverMetrics(auto_pipe_execute=False)
self.subsystem_metrics = s_metrics.Metrics(auto_pipe_execute=False)
self.queue_pop = 0
self.queue_name = settings.CALLBACK_QUEUE
self.prof = AWXProfiler("CallbackBrokerWorker")

View File

@@ -5,7 +5,6 @@
import copy
import json
import re
import sys
import urllib.parse
from jinja2 import sandbox, StrictUndefined
@@ -407,13 +406,11 @@ class SmartFilterField(models.TextField):
# https://docs.python.org/2/library/stdtypes.html#truth-value-testing
if not value:
return None
# avoid doing too much during migrations
if 'migrate' not in sys.argv:
value = urllib.parse.unquote(value)
try:
SmartFilter().query_from_string(value)
except RuntimeError as e:
raise models.base.ValidationError(e)
value = urllib.parse.unquote(value)
try:
SmartFilter().query_from_string(value)
except RuntimeError as e:
raise models.base.ValidationError(e)
return super(SmartFilterField, self).get_prep_value(value)

View File

@@ -1,53 +0,0 @@
# Copyright (c) 2015 Ansible, Inc.
# All Rights Reserved
from django.core.management.base import BaseCommand
from awx.main.models import Instance, ReceptorAddress
def add_address(**kwargs):
try:
instance = Instance.objects.get(hostname=kwargs.pop('instance'))
kwargs['instance'] = instance
if kwargs.get('canonical') and instance.receptor_addresses.filter(canonical=True).exclude(address=kwargs['address']).exists():
print(f"Instance {instance.hostname} already has a canonical address, skipping")
return False
# if ReceptorAddress already exists with address, just update
# otherwise, create new ReceptorAddress
addr, _ = ReceptorAddress.objects.update_or_create(address=kwargs.pop('address'), defaults=kwargs)
print(f"Successfully added receptor address {addr.get_full_address()}")
return True
except Exception as e:
print(f"Error adding receptor address: {e}")
return False
class Command(BaseCommand):
"""
Internal controller command.
Register receptor address to an already-registered instance.
"""
help = "Add receptor address to an instance."
def add_arguments(self, parser):
parser.add_argument('--instance', dest='instance', required=True, type=str, help="Instance hostname this address is added to")
parser.add_argument('--address', dest='address', required=True, type=str, help="Receptor address")
parser.add_argument('--port', dest='port', type=int, help="Receptor listener port")
parser.add_argument('--websocket_path', dest='websocket_path', type=str, default="", help="Path for websockets")
parser.add_argument('--is_internal', action='store_true', help="If true, address only resolvable within the Kubernetes cluster")
parser.add_argument('--protocol', type=str, default='tcp', choices=['tcp', 'ws', 'wss'], help="Protocol to use for the Receptor listener")
parser.add_argument('--canonical', action='store_true', help="If true, address is the canonical address for the instance")
parser.add_argument('--peers_from_control_nodes', action='store_true', help="If true, control nodes will peer to this address")
def handle(self, **options):
address_options = {
k: options[k]
for k in ('instance', 'address', 'port', 'websocket_path', 'is_internal', 'protocol', 'peers_from_control_nodes', 'canonical')
if options[k]
}
changed = add_address(**address_options)
if changed:
print("(changed: True)")

View File

@@ -55,7 +55,7 @@ class Command(BaseCommand):
capacity = f' capacity={x.capacity}' if x.node_type != 'hop' else ''
version = f" version={x.version or '?'}" if x.node_type != 'hop' else ''
heartbeat = f' heartbeat="{x.last_seen:%Y-%m-%d %H:%M:%S}"' if x.last_seen else ''
heartbeat = f' heartbeat="{x.last_seen:%Y-%m-%d %H:%M:%S}"' if x.capacity or x.node_type == 'hop' else ''
print(f'\t{color}{x.hostname}{capacity} node_type={x.node_type}{version}{heartbeat}{end_color}')
print()

View File

@@ -25,17 +25,20 @@ class Command(BaseCommand):
def add_arguments(self, parser):
parser.add_argument('--hostname', dest='hostname', type=str, help="Hostname used during provisioning")
parser.add_argument('--listener_port', dest='listener_port', type=int, help="Receptor listener port")
parser.add_argument('--node_type', type=str, default='hybrid', choices=['control', 'execution', 'hop', 'hybrid'], help="Instance Node type")
parser.add_argument('--uuid', type=str, help="Instance UUID")
def _register_hostname(self, hostname, node_type, uuid):
def _register_hostname(self, hostname, node_type, uuid, listener_port):
if not hostname:
if not settings.AWX_AUTO_DEPROVISION_INSTANCES:
raise CommandError('Registering with values from settings only intended for use in K8s installs')
from awx.main.management.commands.register_queue import RegisterQueue
(changed, instance) = Instance.objects.register(ip_address=os.environ.get('MY_POD_IP'), node_type='control', node_uuid=settings.SYSTEM_UUID)
(changed, instance) = Instance.objects.register(
ip_address=os.environ.get('MY_POD_IP'), listener_port=listener_port, node_type='control', node_uuid=settings.SYSTEM_UUID
)
RegisterQueue(settings.DEFAULT_CONTROL_PLANE_QUEUE_NAME, 100, 0, [], is_container_group=False).register()
RegisterQueue(
settings.DEFAULT_EXECUTION_QUEUE_NAME,
@@ -48,17 +51,16 @@ class Command(BaseCommand):
max_concurrent_jobs=settings.DEFAULT_EXECUTION_QUEUE_MAX_CONCURRENT_JOBS,
).register()
else:
(changed, instance) = Instance.objects.register(hostname=hostname, node_type=node_type, node_uuid=uuid)
(changed, instance) = Instance.objects.register(hostname=hostname, node_type=node_type, node_uuid=uuid, listener_port=listener_port)
if changed:
print("Successfully registered instance {}".format(hostname))
else:
print("Instance already registered {}".format(instance.hostname))
self.changed = changed
@transaction.atomic
def handle(self, **options):
self.changed = False
self._register_hostname(options.get('hostname'), options.get('node_type'), options.get('uuid'))
self._register_hostname(options.get('hostname'), options.get('node_type'), options.get('uuid'), options.get('listener_port'))
if self.changed:
print("(changed: True)")

View File

@@ -1,7 +1,9 @@
import warnings
from django.core.management.base import BaseCommand, CommandError
from django.db import transaction
from awx.main.models import Instance, InstanceLink, ReceptorAddress
from awx.main.models import Instance, InstanceLink
class Command(BaseCommand):
@@ -26,9 +28,7 @@ class Command(BaseCommand):
def handle(self, **options):
# provides a mapping of hostname to Instance objects
nodes = Instance.objects.all().in_bulk(field_name='hostname')
# provides a mapping of address to ReceptorAddress objects
addresses = ReceptorAddress.objects.all().in_bulk(field_name='address')
nodes = Instance.objects.in_bulk(field_name='hostname')
if options['source'] not in nodes:
raise CommandError(f"Host {options['source']} is not a registered instance.")
@@ -39,14 +39,6 @@ class Command(BaseCommand):
if options['exact'] is not None and options['disconnect']:
raise CommandError("The option --disconnect may not be used with --exact.")
# make sure each target has a receptor address
peers = options['peers'] or []
disconnect = options['disconnect'] or []
exact = options['exact'] or []
for peer in peers + disconnect + exact:
if peer not in addresses:
raise CommandError(f"Peer {peer} does not have a receptor address.")
# No 1-cycles
for collection in ('peers', 'disconnect', 'exact'):
if options[collection] is not None and options['source'] in options[collection]:
@@ -55,12 +47,9 @@ class Command(BaseCommand):
# No 2-cycles
if options['peers'] or options['exact'] is not None:
peers = set(options['peers'] or options['exact'])
if options['source'] in addresses:
incoming = set(InstanceLink.objects.filter(target=addresses[options['source']]).values_list('source__hostname', flat=True))
else:
incoming = set()
incoming = set(InstanceLink.objects.filter(target=nodes[options['source']]).values_list('source__hostname', flat=True))
if peers & incoming:
raise CommandError(f"Source node {options['source']} should not link to nodes already peering to it: {peers & incoming}.")
warnings.warn(f"Source node {options['source']} should not link to nodes already peering to it: {peers & incoming}.")
if options['peers']:
missing_peers = set(options['peers']) - set(nodes)
@@ -71,7 +60,7 @@ class Command(BaseCommand):
results = 0
for target in options['peers']:
_, created = InstanceLink.objects.update_or_create(
source=nodes[options['source']], target=addresses[target], defaults={'link_state': InstanceLink.States.ESTABLISHED}
source=nodes[options['source']], target=nodes[target], defaults={'link_state': InstanceLink.States.ESTABLISHED}
)
if created:
results += 1
@@ -81,9 +70,9 @@ class Command(BaseCommand):
if options['disconnect']:
results = 0
for target in options['disconnect']:
if target not in addresses: # Be permissive, the node might have already been de-registered.
if target not in nodes: # Be permissive, the node might have already been de-registered.
continue
n, _ = InstanceLink.objects.filter(source=nodes[options['source']], target=addresses[target]).delete()
n, _ = InstanceLink.objects.filter(source=nodes[options['source']], target=nodes[target]).delete()
results += n
print(f"{results} peer links removed from the database.")
@@ -92,11 +81,11 @@ class Command(BaseCommand):
additions = 0
with transaction.atomic():
peers = set(options['exact'])
links = set(InstanceLink.objects.filter(source=nodes[options['source']]).values_list('target__address', flat=True))
removals, _ = InstanceLink.objects.filter(source=nodes[options['source']], target__instance__hostname__in=links - peers).delete()
links = set(InstanceLink.objects.filter(source=nodes[options['source']]).values_list('target__hostname', flat=True))
removals, _ = InstanceLink.objects.filter(source=nodes[options['source']], target__hostname__in=links - peers).delete()
for target in peers - links:
_, created = InstanceLink.objects.update_or_create(
source=nodes[options['source']], target=addresses[target], defaults={'link_state': InstanceLink.States.ESTABLISHED}
source=nodes[options['source']], target=nodes[target], defaults={'link_state': InstanceLink.States.ESTABLISHED}
)
if created:
additions += 1

View File

@@ -1,26 +0,0 @@
# Copyright (c) 2015 Ansible, Inc.
# All Rights Reserved
from django.core.management.base import BaseCommand
from awx.main.models import ReceptorAddress
class Command(BaseCommand):
"""
Internal controller command.
Delete a receptor address.
"""
help = "Add receptor address to an instance."
def add_arguments(self, parser):
parser.add_argument('--address', dest='address', type=str, help="Receptor address to remove")
def handle(self, **options):
deleted = ReceptorAddress.objects.filter(address=options['address']).delete()
if deleted[0]:
print(f"Successfully removed {options['address']}")
print("(changed: True)")
else:
print(f"Did not remove {options['address']}, not found")

View File

@@ -3,7 +3,6 @@
from django.conf import settings
from django.core.management.base import BaseCommand
from awx.main.analytics.subsystem_metrics import CallbackReceiverMetricsServer
from awx.main.dispatch.control import Control
from awx.main.dispatch.worker import AWXConsumerRedis, CallbackBrokerWorker
@@ -26,9 +25,6 @@ class Command(BaseCommand):
print(Control('callback_receiver').status())
return
consumer = None
CallbackReceiverMetricsServer().start()
try:
consumer = AWXConsumerRedis(
'callback_receiver',

View File

@@ -10,7 +10,6 @@ from awx.main.dispatch import get_task_queuename
from awx.main.dispatch.control import Control
from awx.main.dispatch.pool import AutoscalePool
from awx.main.dispatch.worker import AWXConsumerPG, TaskWorker
from awx.main.analytics.subsystem_metrics import DispatcherMetricsServer
logger = logging.getLogger('awx.main.dispatch')
@@ -63,8 +62,6 @@ class Command(BaseCommand):
consumer = None
DispatcherMetricsServer().start()
try:
queues = ['tower_broadcast_all', 'tower_settings_change', get_task_queuename()]
consumer = AWXConsumerPG('dispatcher', TaskWorker(), queues, AutoscalePool(min_workers=4), schedule=settings.CELERYBEAT_SCHEDULE)

View File

@@ -16,7 +16,6 @@ from awx.main.analytics.broadcast_websocket import (
RelayWebsocketStatsManager,
safe_name,
)
from awx.main.analytics.subsystem_metrics import WebsocketsMetricsServer
from awx.main.wsrelay import WebSocketRelayManager
@@ -92,8 +91,6 @@ class Command(BaseCommand):
return host_stats
def handle(self, *arg, **options):
WebsocketsMetricsServer().start()
# it's necessary to delay this import in case
# database migrations are still running
from awx.main.models.ha import Instance

View File

@@ -115,14 +115,7 @@ class InstanceManager(models.Manager):
return node[0]
raise RuntimeError("No instance found with the current cluster host id")
def register(
self,
node_uuid=None,
hostname=None,
ip_address="",
node_type='hybrid',
defaults=None,
):
def register(self, node_uuid=None, hostname=None, ip_address="", listener_port=None, node_type='hybrid', defaults=None):
if not hostname:
hostname = settings.CLUSTER_HOST_ID
@@ -168,6 +161,9 @@ class InstanceManager(models.Manager):
if instance.node_type != node_type:
instance.node_type = node_type
update_fields.append('node_type')
if instance.listener_port != listener_port:
instance.listener_port = listener_port
update_fields.append('listener_port')
if update_fields:
instance.save(update_fields=update_fields)
return (True, instance)
@@ -178,13 +174,11 @@ class InstanceManager(models.Manager):
create_defaults = {
'node_state': Instance.States.INSTALLED,
'capacity': 0,
'managed': True,
}
if defaults is not None:
create_defaults.update(defaults)
uuid_option = {'uuid': node_uuid if node_uuid is not None else uuid.uuid4()}
if node_type == 'execution' and 'version' not in create_defaults:
create_defaults['version'] = RECEPTOR_PENDING
instance = self.create(hostname=hostname, ip_address=ip_address, node_type=node_type, **create_defaults, **uuid_option)
instance = self.create(hostname=hostname, ip_address=ip_address, listener_port=listener_port, node_type=node_type, **create_defaults, **uuid_option)
return (True, instance)

View File

@@ -1,150 +0,0 @@
# Generated by Django 4.2.6 on 2024-01-19 19:24
import django.core.validators
from django.db import migrations, models
import django.db.models.deletion
def create_receptor_addresses(apps, schema_editor):
"""
If listener_port was defined on an instance, create a receptor address for it
"""
Instance = apps.get_model('main', 'Instance')
ReceptorAddress = apps.get_model('main', 'ReceptorAddress')
for instance in Instance.objects.exclude(listener_port=None):
ReceptorAddress.objects.create(
instance=instance,
address=instance.hostname,
port=instance.listener_port,
peers_from_control_nodes=instance.peers_from_control_nodes,
protocol='tcp',
is_internal=False,
canonical=True,
)
def link_to_receptor_addresses(apps, schema_editor):
"""
Modify each InstanceLink to point to the newly created
ReceptorAddresses, using the new target field
"""
InstanceLink = apps.get_model('main', 'InstanceLink')
for link in InstanceLink.objects.all():
link.target = link.target_old.receptor_addresses.get()
link.save()
class Migration(migrations.Migration):
dependencies = [
('main', '0188_add_bitbucket_dc_webhook'),
]
operations = [
migrations.CreateModel(
name='ReceptorAddress',
fields=[
('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
('address', models.CharField(help_text='Routable address for this instance.', max_length=255)),
(
'port',
models.IntegerField(
default=27199,
help_text='Port for the address.',
validators=[django.core.validators.MinValueValidator(0), django.core.validators.MaxValueValidator(65535)],
),
),
('websocket_path', models.CharField(blank=True, default='', help_text='Websocket path.', max_length=255)),
(
'protocol',
models.CharField(
choices=[('tcp', 'TCP'), ('ws', 'WS'), ('wss', 'WSS')],
default='tcp',
help_text="Protocol to use for the Receptor listener, 'tcp', 'wss', or 'ws'.",
max_length=10,
),
),
('is_internal', models.BooleanField(default=False, help_text='If True, only routable within the Kubernetes cluster.')),
('canonical', models.BooleanField(default=False, help_text='If True, this address is the canonical address for the instance.')),
(
'peers_from_control_nodes',
models.BooleanField(default=False, help_text='If True, control plane cluster nodes should automatically peer to it.'),
),
],
),
migrations.RemoveConstraint(
model_name='instancelink',
name='source_and_target_can_not_be_equal',
),
migrations.RenameField(
model_name='instancelink',
old_name='target',
new_name='target_old',
),
migrations.AlterUniqueTogether(
name='instancelink',
unique_together=set(),
),
migrations.AddField(
model_name='instance',
name='managed',
field=models.BooleanField(default=False, editable=False, help_text='If True, this instance is managed by the control plane.'),
),
migrations.AlterField(
model_name='instancelink',
name='source',
field=models.ForeignKey(help_text='The source instance of this peer link.', on_delete=django.db.models.deletion.CASCADE, to='main.instance'),
),
migrations.AddField(
model_name='receptoraddress',
name='instance',
field=models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, related_name='receptor_addresses', to='main.instance'),
),
migrations.AddField(
model_name='activitystream',
name='receptor_address',
field=models.ManyToManyField(blank=True, to='main.receptoraddress'),
),
migrations.AddConstraint(
model_name='receptoraddress',
constraint=models.UniqueConstraint(fields=('address',), name='unique_receptor_address', violation_error_message='Receptor address must be unique.'),
),
migrations.AddField(
model_name='instancelink',
name='target',
field=models.ForeignKey(
help_text='The target receptor address of this peer link.', null=True, on_delete=django.db.models.deletion.CASCADE, to='main.receptoraddress'
),
),
migrations.RunPython(create_receptor_addresses),
migrations.RunPython(link_to_receptor_addresses),
migrations.RemoveField(
model_name='instance',
name='peers_from_control_nodes',
),
migrations.RemoveField(
model_name='instance',
name='listener_port',
),
migrations.RemoveField(
model_name='instancelink',
name='target_old',
),
migrations.AlterField(
model_name='instance',
name='peers',
field=models.ManyToManyField(related_name='peers_from', through='main.InstanceLink', to='main.receptoraddress'),
),
migrations.AlterField(
model_name='instancelink',
name='target',
field=models.ForeignKey(
help_text='The target receptor address of this peer link.', on_delete=django.db.models.deletion.CASCADE, to='main.receptoraddress'
),
),
migrations.AddConstraint(
model_name='instancelink',
constraint=models.UniqueConstraint(
fields=('source', 'target'), name='unique_source_target', violation_error_message='Field source and target must be unique together.'
),
),
]

View File

@@ -14,7 +14,6 @@ from awx.main.models.unified_jobs import UnifiedJob, UnifiedJobTemplate, StdoutM
from awx.main.models.organization import Organization, Profile, Team, UserSessionMembership # noqa
from awx.main.models.credential import Credential, CredentialType, CredentialInputSource, ManagedCredentialType, build_safe_env # noqa
from awx.main.models.projects import Project, ProjectUpdate # noqa
from awx.main.models.receptor_address import ReceptorAddress # noqa
from awx.main.models.inventory import ( # noqa
CustomInventoryScript,
Group,

View File

@@ -77,7 +77,6 @@ class ActivityStream(models.Model):
notification_template = models.ManyToManyField("NotificationTemplate", blank=True)
notification = models.ManyToManyField("Notification", blank=True)
label = models.ManyToManyField("Label", blank=True)
receptor_address = models.ManyToManyField("ReceptorAddress", blank=True)
role = models.ManyToManyField("Role", blank=True)
instance = models.ManyToManyField("Instance", blank=True)
instance_group = models.ManyToManyField("InstanceGroup", blank=True)

View File

@@ -1216,26 +1216,6 @@ ManagedCredentialType(
},
)
ManagedCredentialType(
namespace='terraform',
kind='cloud',
name=gettext_noop('Terraform backend configuration'),
managed=True,
inputs={
'fields': [
{
'id': 'configuration',
'label': gettext_noop('Backend configuration'),
'type': 'string',
'secret': True,
'multiline': True,
'help_text': gettext_noop('Terraform backend config as Hashicorp configuration language.'),
},
],
'required': ['configuration'],
},
)
class CredentialInputSource(PrimordialModel):
class Meta:

View File

@@ -122,11 +122,3 @@ def kubernetes_bearer_token(cred, env, private_data_dir):
env['K8S_AUTH_SSL_CA_CERT'] = to_container_path(path, private_data_dir)
else:
env['K8S_AUTH_VERIFY_SSL'] = 'False'
def terraform(cred, env, private_data_dir):
handle, path = tempfile.mkstemp(dir=os.path.join(private_data_dir, 'env'))
with os.fdopen(handle, 'w') as f:
os.chmod(path, stat.S_IRUSR | stat.S_IWUSR)
f.write(cred.get_input('configuration'))
env['TF_BACKEND_CONFIG_FILE'] = to_container_path(path, private_data_dir)

View File

@@ -124,6 +124,8 @@ class BasePlaybookEvent(CreatedModifiedModel):
'parent_uuid',
'start_line',
'end_line',
'host_id',
'host_name',
'verbosity',
]
WRAPUP_EVENT = 'playbook_on_stats'
@@ -471,7 +473,7 @@ class JobEvent(BasePlaybookEvent):
An event/message logged from the callback when running a job.
"""
VALID_KEYS = BasePlaybookEvent.VALID_KEYS + ['job_id', 'workflow_job_id', 'job_created', 'host_id', 'host_name']
VALID_KEYS = BasePlaybookEvent.VALID_KEYS + ['job_id', 'workflow_job_id', 'job_created']
JOB_REFERENCE = 'job_id'
objects = DeferJobCreatedManager()

View File

@@ -5,7 +5,7 @@ from decimal import Decimal
import logging
import os
from django.core.validators import MinValueValidator
from django.core.validators import MinValueValidator, MaxValueValidator
from django.db import models, connection
from django.db.models.signals import post_save, post_delete
from django.dispatch import receiver
@@ -34,7 +34,6 @@ from awx.main.models.rbac import (
from awx.main.models.unified_jobs import UnifiedJob
from awx.main.utils.common import get_corrected_cpu, get_cpu_effective_capacity, get_corrected_memory, get_mem_effective_capacity
from awx.main.models.mixins import RelatedJobsMixin, ResourceMixin
from awx.main.models.receptor_address import ReceptorAddress
# ansible-runner
from ansible_runner.utils.capacity import get_cpu_count, get_mem_in_bytes
@@ -65,19 +64,8 @@ class HasPolicyEditsMixin(HasEditsMixin):
class InstanceLink(BaseModel):
class Meta:
ordering = ("id",)
# add constraint for source and target to be unique together
constraints = [
models.UniqueConstraint(
fields=["source", "target"],
name="unique_source_target",
violation_error_message=_("Field source and target must be unique together."),
)
]
source = models.ForeignKey('Instance', on_delete=models.CASCADE, help_text=_("The source instance of this peer link."))
target = models.ForeignKey('ReceptorAddress', on_delete=models.CASCADE, help_text=_("The target receptor address of this peer link."))
source = models.ForeignKey('Instance', on_delete=models.CASCADE, related_name='+')
target = models.ForeignKey('Instance', on_delete=models.CASCADE, related_name='reverse_peers')
class States(models.TextChoices):
ADDING = 'adding', _('Adding')
@@ -88,6 +76,11 @@ class InstanceLink(BaseModel):
choices=States.choices, default=States.ADDING, max_length=16, help_text=_("Indicates the current life cycle stage of this peer link.")
)
class Meta:
unique_together = ('source', 'target')
ordering = ("id",)
constraints = [models.CheckConstraint(check=~models.Q(source=models.F('target')), name='source_and_target_can_not_be_equal')]
class Instance(HasPolicyEditsMixin, BaseModel):
"""A model representing an AWX instance running against this database."""
@@ -117,7 +110,6 @@ class Instance(HasPolicyEditsMixin, BaseModel):
default="",
max_length=50,
)
# Auto-fields, implementation is different from BaseModel
created = models.DateTimeField(auto_now_add=True)
modified = models.DateTimeField(auto_now=True)
@@ -193,9 +185,16 @@ class Instance(HasPolicyEditsMixin, BaseModel):
node_state = models.CharField(
choices=States.choices, default=States.READY, max_length=16, help_text=_("Indicates the current life cycle stage of this instance.")
)
listener_port = models.PositiveIntegerField(
blank=True,
null=True,
default=None,
validators=[MinValueValidator(1024), MaxValueValidator(65535)],
help_text=_("Port that Receptor will listen for incoming connections on."),
)
managed = models.BooleanField(help_text=_("If True, this instance is managed by the control plane."), default=False, editable=False)
peers = models.ManyToManyField('ReceptorAddress', through=InstanceLink, through_fields=('source', 'target'), related_name='peers_from')
peers = models.ManyToManyField('self', symmetrical=False, through=InstanceLink, through_fields=('source', 'target'), related_name='peers_from')
peers_from_control_nodes = models.BooleanField(default=False, help_text=_("If True, control plane cluster nodes should automatically peer to it."))
POLICY_FIELDS = frozenset(('managed_by_policy', 'hostname', 'capacity_adjustment'))
@@ -242,26 +241,6 @@ class Instance(HasPolicyEditsMixin, BaseModel):
return True
return self.health_check_started > self.last_health_check
@property
def canonical_address(self):
return self.receptor_addresses.filter(canonical=True).first()
@property
def canonical_address_port(self):
# note: don't create a different query for receptor addresses, as this is prefetched on the View for optimization
for addr in self.receptor_addresses.all():
if addr.canonical:
return addr.port
return None
@property
def canonical_address_peers_from_control_nodes(self):
# note: don't create a different query for receptor addresses, as this is prefetched on the View for optimization
for addr in self.receptor_addresses.all():
if addr.canonical:
return addr.peers_from_control_nodes
return False
def get_cleanup_task_kwargs(self, **kwargs):
"""
Produce options to use for the command: ansible-runner worker cleanup
@@ -522,35 +501,6 @@ def schedule_write_receptor_config(broadcast=True):
write_receptor_config() # just run locally
@receiver(post_save, sender=ReceptorAddress)
def receptor_address_saved(sender, instance, **kwargs):
from awx.main.signals import disable_activity_stream
address = instance
control_instances = set(Instance.objects.filter(node_type__in=[Instance.Types.CONTROL, Instance.Types.HYBRID]))
if address.peers_from_control_nodes:
# if control_instances is not a subset of current peers of address, then
# that means we need to add some InstanceLinks
if not control_instances <= set(address.peers_from.all()):
with disable_activity_stream():
for control_instance in control_instances:
InstanceLink.objects.update_or_create(source=control_instance, target=address)
schedule_write_receptor_config()
else:
if address.peers_from.exists():
with disable_activity_stream():
address.peers_from.remove(*control_instances)
schedule_write_receptor_config()
@receiver(post_delete, sender=ReceptorAddress)
def receptor_address_deleted(sender, instance, **kwargs):
address = instance
if address.peers_from_control_nodes:
schedule_write_receptor_config()
@receiver(post_save, sender=Instance)
def on_instance_saved(sender, instance, created=False, raw=False, **kwargs):
'''
@@ -561,14 +511,11 @@ def on_instance_saved(sender, instance, created=False, raw=False, **kwargs):
2. a node changes its value of peers_from_control_nodes
3. a new control node comes online and has instances to peer to
'''
from awx.main.signals import disable_activity_stream
if created and settings.IS_K8S and instance.node_type in [Instance.Types.CONTROL, Instance.Types.HYBRID]:
peers_addresses = ReceptorAddress.objects.filter(peers_from_control_nodes=True)
if peers_addresses.exists():
with disable_activity_stream():
instance.peers.add(*peers_addresses)
schedule_write_receptor_config(broadcast=False)
inst = Instance.objects.filter(peers_from_control_nodes=True)
if set(instance.peers.all()) != set(inst):
instance.peers.set(inst)
schedule_write_receptor_config(broadcast=False)
if settings.IS_K8S and instance.node_type in [Instance.Types.HOP, Instance.Types.EXECUTION]:
if instance.node_state == Instance.States.DEPROVISIONING:
@@ -577,6 +524,16 @@ def on_instance_saved(sender, instance, created=False, raw=False, **kwargs):
# wait for jobs on the node to complete, then delete the
# node and kick off write_receptor_config
connection.on_commit(lambda: remove_deprovisioned_node.apply_async([instance.hostname]))
else:
control_instances = set(Instance.objects.filter(node_type__in=[Instance.Types.CONTROL, Instance.Types.HYBRID]))
if instance.peers_from_control_nodes:
if (control_instances & set(instance.peers_from.all())) != set(control_instances):
instance.peers_from.add(*control_instances)
schedule_write_receptor_config() # keep method separate to make pytest mocking easier
else:
if set(control_instances) & set(instance.peers_from.all()):
instance.peers_from.remove(*control_instances)
schedule_write_receptor_config()
if created or instance.has_policy_changes():
schedule_policy_task()
@@ -591,6 +548,8 @@ def on_instance_group_deleted(sender, instance, using, **kwargs):
@receiver(post_delete, sender=Instance)
def on_instance_deleted(sender, instance, using, **kwargs):
schedule_policy_task()
if settings.IS_K8S and instance.node_type in (Instance.Types.EXECUTION, Instance.Types.HOP) and instance.peers_from_control_nodes:
schedule_write_receptor_config()
class UnifiedJobTemplateInstanceGroupMembership(models.Model):

View File

@@ -5,7 +5,6 @@ from copy import deepcopy
import datetime
import logging
import json
import traceback
from django.db import models
from django.conf import settings
@@ -485,29 +484,14 @@ class JobNotificationMixin(object):
if msg_template:
try:
msg = env.from_string(msg_template).render(**context)
except (TemplateSyntaxError, UndefinedError, SecurityError) as e:
msg = '\r\n'.join([e.message, ''.join(traceback.format_exception(None, e, e.__traceback__).replace('\n', '\r\n'))])
except (TemplateSyntaxError, UndefinedError, SecurityError):
msg = ''
if body_template:
try:
body = env.from_string(body_template).render(**context)
except (TemplateSyntaxError, UndefinedError, SecurityError) as e:
body = '\r\n'.join([e.message, ''.join(traceback.format_exception(None, e, e.__traceback__).replace('\n', '\r\n'))])
# https://datatracker.ietf.org/doc/html/rfc2822#section-2.2
# Body should have at least 2 CRLF, some clients will interpret
# the email incorrectly with blank body. So we will check that
if len(body.strip().splitlines()) <= 2:
# blank body
body = '\r\n'.join(
[
"The template rendering return a blank body.",
"Please check the template.",
"Refer to https://github.com/ansible/awx/issues/13983",
"for further information.",
]
)
except (TemplateSyntaxError, UndefinedError, SecurityError):
body = ''
return (msg, body)

View File

@@ -1,67 +0,0 @@
from django.db import models
from django.core.validators import MinValueValidator, MaxValueValidator
from django.utils.translation import gettext_lazy as _
from awx.api.versioning import reverse
class Protocols(models.TextChoices):
TCP = 'tcp', 'TCP'
WS = 'ws', 'WS'
WSS = 'wss', 'WSS'
class ReceptorAddress(models.Model):
class Meta:
app_label = 'main'
constraints = [
models.UniqueConstraint(
fields=["address"],
name="unique_receptor_address",
violation_error_message=_("Receptor address must be unique."),
)
]
address = models.CharField(help_text=_("Routable address for this instance."), max_length=255)
port = models.IntegerField(help_text=_("Port for the address."), default=27199, validators=[MinValueValidator(0), MaxValueValidator(65535)])
websocket_path = models.CharField(help_text=_("Websocket path."), max_length=255, default="", blank=True)
protocol = models.CharField(
help_text=_("Protocol to use for the Receptor listener, 'tcp', 'wss', or 'ws'."), max_length=10, default=Protocols.TCP, choices=Protocols.choices
)
is_internal = models.BooleanField(help_text=_("If True, only routable within the Kubernetes cluster."), default=False)
canonical = models.BooleanField(help_text=_("If True, this address is the canonical address for the instance."), default=False)
peers_from_control_nodes = models.BooleanField(help_text=_("If True, control plane cluster nodes should automatically peer to it."), default=False)
instance = models.ForeignKey(
'Instance',
related_name='receptor_addresses',
on_delete=models.CASCADE,
null=False,
)
def __str__(self):
return self.get_full_address()
def get_full_address(self):
scheme = ""
path = ""
port = ""
if self.protocol == "ws":
scheme = "wss://"
if self.protocol == "ws" and self.websocket_path:
path = f"/{self.websocket_path}"
if self.port:
port = f":{self.port}"
return f"{scheme}{self.address}{port}{path}"
def get_peer_type(self):
if self.protocol == 'tcp':
return 'tcp-peer'
elif self.protocol in ['ws', 'wss']:
return 'ws-peer'
else:
return None
def get_absolute_url(self, request=None):
return reverse('api:receptor_address_detail', kwargs={'pk': self.pk}, request=request)

View File

@@ -4,10 +4,9 @@ import logging
from django.conf import settings
from django.urls import re_path
from channels.auth import AuthMiddlewareStack
from channels.routing import ProtocolTypeRouter, URLRouter
from ansible_base.lib.channels.middleware import DrfAuthMiddlewareStack
from . import consumers
@@ -27,50 +26,13 @@ class AWXProtocolTypeRouter(ProtocolTypeRouter):
super().__init__(*args, **kwargs)
class MultipleURLRouterAdapter:
"""
Django channels doesn't nicely support Auth_1(urls_1), Auth_2(urls_2), ..., Auth_n(urls_n)
This class allows assocating a websocket url with an auth
Ordering matters. The first matching url will be used.
"""
def __init__(self, *auths):
self._auths = [a for a in auths]
async def __call__(self, scope, receive, send):
"""
Loop through the list of passed in URLRouter's (they may or may not be wrapped by auth).
We know we have exhausted the list of URLRouter patterns when we get a
ValueError('No route found for path %s'). When that happens, move onto the next
URLRouter.
If the final URLRouter raises an error, re-raise it in the end.
We know that we found a match when no error is raised, end the loop.
"""
last_index = len(self._auths) - 1
for i, auth in enumerate(self._auths):
try:
return await auth.__call__(scope, receive, send)
except ValueError as e:
if str(e).startswith('No route found for path'):
# Only surface the error if on the last URLRouter
if i == last_index:
raise
websocket_urlpatterns = [
re_path(r'api/websocket/$', consumers.EventConsumer.as_asgi()),
re_path(r'websocket/$', consumers.EventConsumer.as_asgi()),
]
websocket_relay_urlpatterns = [
re_path(r'websocket/relay/$', consumers.RelayConsumer.as_asgi()),
]
application = AWXProtocolTypeRouter(
{
'websocket': MultipleURLRouterAdapter(
URLRouter(websocket_relay_urlpatterns),
DrfAuthMiddlewareStack(URLRouter(websocket_urlpatterns)),
)
'websocket': AuthMiddlewareStack(URLRouter(websocket_urlpatterns)),
}
)

View File

@@ -68,7 +68,7 @@ class TaskBase:
# initialize each metric to 0 and force metric_has_changed to true. This
# ensures each task manager metric will be overridden when pipe_execute
# is called later.
self.subsystem_metrics = s_metrics.DispatcherMetrics(auto_pipe_execute=False)
self.subsystem_metrics = s_metrics.Metrics(auto_pipe_execute=False)
self.start_time = time.time()
# We want to avoid calling settings in loops, so cache these settings at init time
@@ -105,7 +105,7 @@ class TaskBase:
try:
# increment task_manager_schedule_calls regardless if the other
# metrics are recorded
s_metrics.DispatcherMetrics(auto_pipe_execute=True).inc(f"{self.prefix}__schedule_calls", 1)
s_metrics.Metrics(auto_pipe_execute=True).inc(f"{self.prefix}__schedule_calls", 1)
# Only record metrics if the last time recording was more
# than SUBSYSTEM_METRICS_TASK_MANAGER_RECORD_INTERVAL ago.
# Prevents a short-duration task manager that runs directly after a

View File

@@ -29,7 +29,7 @@ class RunnerCallback:
self.safe_env = {}
self.event_ct = 0
self.model = model
self.update_attempts = int(getattr(settings, 'DISPATCHER_DB_DOWNTOWN_TOLLERANCE', settings.DISPATCHER_DB_DOWNTIME_TOLERANCE) / 5)
self.update_attempts = int(settings.DISPATCHER_DB_DOWNTIME_TOLERANCE / 5)
self.wrapup_event_dispatched = False
self.artifacts_processed = False
self.extra_update_fields = {}
@@ -95,17 +95,17 @@ class RunnerCallback:
if self.parent_workflow_job_id:
event_data['workflow_job_id'] = self.parent_workflow_job_id
event_data['job_created'] = self.job_created
host = event_data.get('event_data', {}).get('host', '').strip()
if host:
event_data['host_name'] = host
if host in self.host_map:
event_data['host_id'] = self.host_map[host]
else:
event_data['host_name'] = ''
event_data['host_id'] = ''
if event_data.get('event') == 'playbook_on_stats':
event_data['host_map'] = self.host_map
if self.host_map:
host = event_data.get('event_data', {}).get('host', '').strip()
if host:
event_data['host_name'] = host
if host in self.host_map:
event_data['host_id'] = self.host_map[host]
else:
event_data['host_name'] = ''
event_data['host_id'] = ''
if event_data.get('event') == 'playbook_on_stats':
event_data['host_map'] = self.host_map
if isinstance(self, RunnerCallbackForProjectUpdate):
# need a better way to have this check.

View File

@@ -114,7 +114,7 @@ class BaseTask(object):
def __init__(self):
self.cleanup_paths = []
self.update_attempts = int(getattr(settings, 'DISPATCHER_DB_DOWNTOWN_TOLLERANCE', settings.DISPATCHER_DB_DOWNTIME_TOLERANCE) / 5)
self.update_attempts = int(settings.DISPATCHER_DB_DOWNTIME_TOLERANCE / 5)
self.runner_callback = self.callback_class(model=self.model)
def update_model(self, pk, _attempt=0, **updates):

View File

@@ -27,7 +27,7 @@ from awx.main.utils.common import (
)
from awx.main.constants import MAX_ISOLATED_PATH_COLON_DELIMITER
from awx.main.tasks.signals import signal_state, signal_callback, SignalExit
from awx.main.models import Instance, InstanceLink, UnifiedJob, ReceptorAddress
from awx.main.models import Instance, InstanceLink, UnifiedJob
from awx.main.dispatch import get_task_queuename
from awx.main.dispatch.publish import task
from awx.main.utils.pglock import advisory_lock
@@ -676,44 +676,36 @@ RECEPTOR_CONFIG_STARTER = (
)
def should_update_config(new_config):
def should_update_config(instances):
'''
checks that the list of instances matches the list of
tcp-peers in the config
'''
current_config = read_receptor_config() # this gets receptor conf lock
current_peers = []
for config_entry in current_config:
if config_entry not in new_config:
logger.warning(f"{config_entry} should not be in receptor config. Updating.")
return True
for config_entry in new_config:
if config_entry not in current_config:
logger.warning(f"{config_entry} missing from receptor config. Updating.")
return True
for key, value in config_entry.items():
if key.endswith('-peer'):
current_peers.append(value['address'])
intended_peers = [f"{i.hostname}:{i.listener_port}" for i in instances]
logger.debug(f"Peers current {current_peers} intended {intended_peers}")
if set(current_peers) == set(intended_peers):
return False # config file is already update to date
return False
return True
def generate_config_data():
# returns two values
# receptor config - based on current database peers
# should_update - If True, receptor_config differs from the receptor conf file on disk
addresses = ReceptorAddress.objects.filter(peers_from_control_nodes=True)
instances = Instance.objects.filter(node_type__in=(Instance.Types.EXECUTION, Instance.Types.HOP), peers_from_control_nodes=True)
receptor_config = list(RECEPTOR_CONFIG_STARTER)
for address in addresses:
if address.get_peer_type():
peer = {
f'{address.get_peer_type()}': {
'address': f'{address.get_full_address()}',
'tls': 'tlsclient',
}
}
receptor_config.append(peer)
else:
logger.warning(f"Receptor address {address} has unsupported peer type, skipping.")
should_update = should_update_config(receptor_config)
for instance in instances:
peer = {'tcp-peer': {'address': f'{instance.hostname}:{instance.listener_port}', 'tls': 'tlsclient'}}
receptor_config.append(peer)
should_update = should_update_config(instances)
return receptor_config, should_update
@@ -755,13 +747,14 @@ def write_receptor_config():
with lock:
with open(__RECEPTOR_CONF, 'w') as file:
yaml.dump(receptor_config, file, default_flow_style=False)
reload_receptor()
@task(queue=get_task_queuename)
def remove_deprovisioned_node(hostname):
InstanceLink.objects.filter(source__hostname=hostname).update(link_state=InstanceLink.States.REMOVING)
InstanceLink.objects.filter(target__instance__hostname=hostname).update(link_state=InstanceLink.States.REMOVING)
InstanceLink.objects.filter(target__hostname=hostname).update(link_state=InstanceLink.States.REMOVING)
node_jobs = UnifiedJob.objects.filter(
execution_node=hostname,

View File

@@ -62,7 +62,7 @@ from awx.main.tasks.receptor import get_receptor_ctl, worker_info, worker_cleanu
from awx.main.consumers import emit_channel_notification
from awx.main import analytics
from awx.conf import settings_registry
from awx.main.analytics.subsystem_metrics import DispatcherMetrics
from awx.main.analytics.subsystem_metrics import Metrics
from rest_framework.exceptions import PermissionDenied
@@ -113,7 +113,7 @@ def dispatch_startup():
cluster_node_heartbeat()
reaper.startup_reaping()
reaper.reap_waiting(grace_period=0)
m = DispatcherMetrics()
m = Metrics()
m.reset_values()
@@ -495,7 +495,7 @@ def inspect_established_receptor_connections(mesh_status):
update_links = []
for link in all_links:
if link.link_state != InstanceLink.States.REMOVING:
if link.target.instance.hostname in active_receptor_conns.get(link.source.hostname, {}):
if link.target.hostname in active_receptor_conns.get(link.source.hostname, {}):
if link.link_state is not InstanceLink.States.ESTABLISHED:
link.link_state = InstanceLink.States.ESTABLISHED
update_links.append(link)

View File

@@ -1,16 +1,19 @@
import pytest
import yaml
import itertools
from unittest import mock
from django.db.utils import IntegrityError
from awx.api.versioning import reverse
from awx.main.models import Instance, ReceptorAddress
from awx.main.models import Instance
from awx.api.views.instance_install_bundle import generate_group_vars_all_yml
def has_peer(group_vars, peer):
peers = group_vars.get('receptor_peers', [])
for p in peers:
if p['address'] == peer:
if f"{p['host']}:{p['port']}" == peer:
return True
return False
@@ -21,314 +24,119 @@ class TestPeers:
def configure_settings(self, settings):
settings.IS_K8S = True
@pytest.mark.parametrize('node_type', ['hop', 'execution'])
def test_peering_to_self(self, node_type, admin_user, patch):
@pytest.mark.parametrize('node_type', ['control', 'hybrid'])
def test_prevent_peering_to_self(self, node_type):
"""
cannot peer to self
"""
instance = Instance.objects.create(hostname='abc', node_type=node_type)
addr = ReceptorAddress.objects.create(instance=instance, address='abc', canonical=True)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': instance.pk}),
data={"hostname": "abc", "node_type": node_type, "peers": [addr.id]},
user=admin_user,
expect=400,
)
assert 'Instance cannot peer to its own address.' in str(resp.data)
control_instance = Instance.objects.create(hostname='abc', node_type=node_type)
with pytest.raises(IntegrityError):
control_instance.peers.add(control_instance)
@pytest.mark.parametrize('node_type', ['control', 'hybrid', 'hop', 'execution'])
def test_creating_node(self, node_type, admin_user, post):
"""
can only add hop and execution nodes via API
"""
resp = post(
post(
url=reverse('api:instance_list'),
data={"hostname": "abc", "node_type": node_type},
user=admin_user,
expect=400 if node_type in ['control', 'hybrid'] else 201,
)
if resp.status_code == 400:
assert 'Can only create execution or hop nodes.' in str(resp.data)
def test_changing_node_type(self, admin_user, patch):
"""
cannot change node type
"""
hop = Instance.objects.create(hostname='abc', node_type="hop")
resp = patch(
patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"node_type": "execution"},
user=admin_user,
expect=400,
)
assert 'Cannot change node type.' in str(resp.data)
@pytest.mark.parametrize(
'payload_port, payload_peers_from, initial_port, initial_peers_from',
[
(-1, -1, None, None),
(-1, -1, 27199, False),
(-1, -1, 27199, True),
(None, -1, None, None),
(None, False, None, None),
(-1, False, None, None),
(27199, True, 27199, True),
(27199, False, 27199, False),
(27199, -1, 27199, True),
(27199, -1, 27199, False),
(-1, True, 27199, True),
(-1, False, 27199, False),
],
)
def test_no_op(self, payload_port, payload_peers_from, initial_port, initial_peers_from, admin_user, patch):
node = Instance.objects.create(hostname='abc', node_type='hop')
if initial_port is not None:
ReceptorAddress.objects.create(address=node.hostname, port=initial_port, canonical=True, peers_from_control_nodes=initial_peers_from, instance=node)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
else:
assert ReceptorAddress.objects.filter(instance=node).count() == 0
data = {'enabled': True} # Just to have something to post.
if payload_port != -1:
data['listener_port'] = payload_port
if payload_peers_from != -1:
data['peers_from_control_nodes'] = payload_peers_from
patch(
url=reverse('api:instance_detail', kwargs={'pk': node.pk}),
data=data,
@pytest.mark.parametrize('node_type', ['hop', 'execution'])
def test_listener_port_null(self, node_type, admin_user, post):
"""
listener_port can be None
"""
post(
url=reverse('api:instance_list'),
data={"hostname": "abc", "node_type": node_type, "listener_port": None},
user=admin_user,
expect=200,
expect=201,
)
assert ReceptorAddress.objects.filter(instance=node).count() == (0 if initial_port is None else 1)
if initial_port is not None:
ra = ReceptorAddress.objects.get(instance=node, canonical=True)
assert ra.port == initial_port
assert ra.peers_from_control_nodes == initial_peers_from
@pytest.mark.parametrize(
'payload_port, payload_peers_from',
[
(27199, True),
(27199, False),
(27199, -1),
],
)
def test_creates_canonical_address(self, payload_port, payload_peers_from, admin_user, patch):
node = Instance.objects.create(hostname='abc', node_type='hop')
assert ReceptorAddress.objects.filter(instance=node).count() == 0
data = {'enabled': True} # Just to have something to post.
if payload_port != -1:
data['listener_port'] = payload_port
if payload_peers_from != -1:
data['peers_from_control_nodes'] = payload_peers_from
patch(
url=reverse('api:instance_detail', kwargs={'pk': node.pk}),
data=data,
@pytest.mark.parametrize('node_type, allowed', [('control', False), ('hybrid', False), ('hop', True), ('execution', True)])
def test_peers_from_control_nodes_allowed(self, node_type, allowed, post, admin_user):
"""
only hop and execution nodes can have peers_from_control_nodes set to True
"""
post(
url=reverse('api:instance_list'),
data={"hostname": "abc", "peers_from_control_nodes": True, "node_type": node_type, "listener_port": 6789},
user=admin_user,
expect=200,
expect=201 if allowed else 400,
)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
ra = ReceptorAddress.objects.get(instance=node, canonical=True)
assert ra.port == payload_port
assert ra.peers_from_control_nodes == (payload_peers_from if payload_peers_from != -1 else False)
@pytest.mark.parametrize(
'payload_port, payload_peers_from, initial_port, initial_peers_from',
[
(None, False, 27199, True),
(None, -1, 27199, True),
(None, False, 27199, False),
(None, -1, 27199, False),
],
)
def test_deletes_canonical_address(self, payload_port, payload_peers_from, initial_port, initial_peers_from, admin_user, patch):
node = Instance.objects.create(hostname='abc', node_type='hop')
ReceptorAddress.objects.create(address=node.hostname, port=initial_port, canonical=True, peers_from_control_nodes=initial_peers_from, instance=node)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
data = {'enabled': True} # Just to have something to post.
if payload_port != -1:
data['listener_port'] = payload_port
if payload_peers_from != -1:
data['peers_from_control_nodes'] = payload_peers_from
patch(
url=reverse('api:instance_detail', kwargs={'pk': node.pk}),
data=data,
user=admin_user,
expect=200,
)
assert ReceptorAddress.objects.filter(instance=node).count() == 0
@pytest.mark.parametrize(
'payload_port, payload_peers_from, initial_port, initial_peers_from',
[
(27199, True, 27199, False),
(27199, False, 27199, True),
(-1, True, 27199, False),
(-1, False, 27199, True),
],
)
def test_updates_canonical_address(self, payload_port, payload_peers_from, initial_port, initial_peers_from, admin_user, patch):
node = Instance.objects.create(hostname='abc', node_type='hop')
ReceptorAddress.objects.create(address=node.hostname, port=initial_port, canonical=True, peers_from_control_nodes=initial_peers_from, instance=node)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
data = {'enabled': True} # Just to have something to post.
if payload_port != -1:
data['listener_port'] = payload_port
if payload_peers_from != -1:
data['peers_from_control_nodes'] = payload_peers_from
patch(
url=reverse('api:instance_detail', kwargs={'pk': node.pk}),
data=data,
user=admin_user,
expect=200,
)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
ra = ReceptorAddress.objects.get(instance=node, canonical=True)
assert ra.port == initial_port # At the present time, changing ports is not allowed
assert ra.peers_from_control_nodes == payload_peers_from
@pytest.mark.parametrize(
'payload_port, payload_peers_from, initial_port, initial_peers_from, error_msg',
[
(-1, True, None, None, "Cannot enable peers_from_control_nodes"),
(None, True, None, None, "Cannot enable peers_from_control_nodes"),
(None, True, 21799, True, "Cannot enable peers_from_control_nodes"),
(None, True, 21799, False, "Cannot enable peers_from_control_nodes"),
(21800, -1, 21799, True, "Cannot change listener port"),
(21800, True, 21799, True, "Cannot change listener port"),
(21800, False, 21799, True, "Cannot change listener port"),
(21800, -1, 21799, False, "Cannot change listener port"),
(21800, True, 21799, False, "Cannot change listener port"),
(21800, False, 21799, False, "Cannot change listener port"),
],
)
def test_canonical_address_validation_error(self, payload_port, payload_peers_from, initial_port, initial_peers_from, error_msg, admin_user, patch):
node = Instance.objects.create(hostname='abc', node_type='hop')
if initial_port is not None:
ReceptorAddress.objects.create(address=node.hostname, port=initial_port, canonical=True, peers_from_control_nodes=initial_peers_from, instance=node)
assert ReceptorAddress.objects.filter(instance=node).count() == 1
else:
assert ReceptorAddress.objects.filter(instance=node).count() == 0
data = {'enabled': True} # Just to have something to post.
if payload_port != -1:
data['listener_port'] = payload_port
if payload_peers_from != -1:
data['peers_from_control_nodes'] = payload_peers_from
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': node.pk}),
data=data,
def test_listener_port_is_required(self, admin_user, post):
"""
if adding instance to peers list, that instance must have listener_port set
"""
Instance.objects.create(hostname='abc', node_type="hop", listener_port=None)
post(
url=reverse('api:instance_list'),
data={"hostname": "ex", "peers_from_control_nodes": False, "node_type": "execution", "listener_port": None, "peers": ["abc"]},
user=admin_user,
expect=400,
)
assert error_msg in str(resp.data)
def test_changing_managed_listener_port(self, admin_user, patch):
def test_peers_from_control_nodes_listener_port_enabled(self, admin_user, post):
"""
if instance is managed, cannot change listener port at all
if peers_from_control_nodes is True, listener_port must an integer
Assert that all other combinations are allowed
"""
hop = Instance.objects.create(hostname='abc', node_type="hop", managed=True)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"listener_port": 5678},
user=admin_user,
expect=400, # cannot set port
)
assert 'Cannot change listener port for managed nodes.' in str(resp.data)
ReceptorAddress.objects.create(instance=hop, address='hop', port=27199, canonical=True)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"listener_port": None},
user=admin_user,
expect=400, # cannot unset port
)
assert 'Cannot change listener port for managed nodes.' in str(resp.data)
def test_bidirectional_peering(self, admin_user, patch):
"""
cannot peer to node that is already to peered to it
if A -> B, then disallow B -> A
"""
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop1addr = ReceptorAddress.objects.create(instance=hop1, address='hop1', canonical=True)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', canonical=True)
hop1.peers.add(hop2addr)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop2.pk}),
data={"peers": [hop1addr.id]},
user=admin_user,
expect=400,
)
assert 'Instance hop1 is already peered to this instance.' in str(resp.data)
def test_multiple_peers_same_instance(self, admin_user, patch):
"""
cannot peer to more than one address of the same instance
"""
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop1addr1 = ReceptorAddress.objects.create(instance=hop1, address='hop1', canonical=True)
hop1addr2 = ReceptorAddress.objects.create(instance=hop1, address='hop1alternate')
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop2.pk}),
data={"peers": [hop1addr1.id, hop1addr2.id]},
user=admin_user,
expect=400,
)
assert 'Cannot peer to the same instance more than once.' in str(resp.data)
for index, item in enumerate(itertools.product(['hop', 'execution'], [True, False], [None, 6789])):
node_type, peers_from, listener_port = item
# only disallowed case is when peers_from is True and listener port is None
disallowed = peers_from and not listener_port
post(
url=reverse('api:instance_list'),
data={"hostname": f"abc{index}", "peers_from_control_nodes": peers_from, "node_type": node_type, "listener_port": listener_port},
user=admin_user,
expect=400 if disallowed else 201,
)
@pytest.mark.parametrize('node_type', ['control', 'hybrid'])
def test_changing_peers_control_nodes(self, node_type, admin_user, patch):
def test_disallow_modifying_peers_control_nodes(self, node_type, admin_user, patch):
"""
for control nodes, peers field should not be
modified directly via patch.
"""
control = Instance.objects.create(hostname='abc', node_type=node_type, managed=True)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop1addr = ReceptorAddress.objects.create(instance=hop1, address='hop1', peers_from_control_nodes=True, canonical=True)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', canonical=True)
assert [hop1addr] == list(control.peers.all()) # only hop1addr should be peered
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': control.pk}),
data={"peers": [hop2addr.id]},
user=admin_user,
expect=400, # cannot add peers manually
)
assert 'Setting peers manually for managed nodes is not allowed.' in str(resp.data)
control = Instance.objects.create(hostname='abc', node_type=node_type)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop', peers_from_control_nodes=True, listener_port=6789)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop', peers_from_control_nodes=False, listener_port=6789)
assert [hop1] == list(control.peers.all()) # only hop1 should be peered
patch(
url=reverse('api:instance_detail', kwargs={'pk': control.pk}),
data={"peers": [hop1addr.id]},
data={"peers": ["hop2"]},
user=admin_user,
expect=400, # cannot add peers directly
)
patch(
url=reverse('api:instance_detail', kwargs={'pk': control.pk}),
data={"peers": ["hop1"]},
user=admin_user,
expect=200, # patching with current peers list should be okay
)
resp = patch(
patch(
url=reverse('api:instance_detail', kwargs={'pk': control.pk}),
data={"peers": []},
user=admin_user,
expect=400, # cannot remove peers directly
)
assert 'Setting peers manually for managed nodes is not allowed.' in str(resp.data)
patch(
url=reverse('api:instance_detail', kwargs={'pk': control.pk}),
data={},
@@ -340,25 +148,23 @@ class TestPeers:
url=reverse('api:instance_detail', kwargs={'pk': hop2.pk}),
data={"peers_from_control_nodes": True},
user=admin_user,
expect=200,
expect=200, # patching without data should be fine too
)
assert {hop1addr, hop2addr} == set(control.peers.all()) # hop1 and hop2 should now be peered from control node
assert {hop1, hop2} == set(control.peers.all()) # hop1 and hop2 should now be peered from control node
def test_changing_hostname(self, admin_user, patch):
def test_disallow_changing_hostname(self, admin_user, patch):
"""
cannot change hostname
"""
hop = Instance.objects.create(hostname='hop', node_type='hop')
resp = patch(
patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"hostname": "hop2"},
user=admin_user,
expect=400,
)
assert 'Cannot change hostname.' in str(resp.data)
def test_changing_node_state(self, admin_user, patch):
def test_disallow_changing_node_state(self, admin_user, patch):
"""
only allow setting to deprovisioning
"""
@@ -369,54 +175,12 @@ class TestPeers:
user=admin_user,
expect=200,
)
resp = patch(
patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"node_state": "ready"},
user=admin_user,
expect=400,
)
assert "Can only change instances to the 'deprovisioning' state." in str(resp.data)
def test_changing_managed_node_state(self, admin_user, patch):
"""
cannot change node state of managed node
"""
hop = Instance.objects.create(hostname='hop', node_type='hop', managed=True)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"node_state": "deprovisioning"},
user=admin_user,
expect=400,
)
assert 'Cannot deprovision managed nodes.' in str(resp.data)
def test_changing_managed_peers_from_control_nodes(self, admin_user, patch):
"""
cannot change peers_from_control_nodes of managed node
"""
hop = Instance.objects.create(hostname='hop', node_type='hop', managed=True)
ReceptorAddress.objects.create(instance=hop, address='hop', peers_from_control_nodes=True, canonical=True)
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"peers_from_control_nodes": False},
user=admin_user,
expect=400,
)
assert 'Cannot change peers_from_control_nodes for managed nodes.' in str(resp.data)
hop.peers_from_control_nodes = False
hop.save()
resp = patch(
url=reverse('api:instance_detail', kwargs={'pk': hop.pk}),
data={"peers_from_control_nodes": False},
user=admin_user,
expect=400,
)
assert 'Cannot change peers_from_control_nodes for managed nodes.' in str(resp.data)
@pytest.mark.parametrize('node_type', ['control', 'hybrid'])
def test_control_node_automatically_peers(self, node_type):
@@ -427,10 +191,9 @@ class TestPeers:
peer to hop should be removed if hop is deleted
"""
hop = Instance.objects.create(hostname='hop', node_type='hop')
hopaddr = ReceptorAddress.objects.create(instance=hop, address='hop', peers_from_control_nodes=True, canonical=True)
hop = Instance.objects.create(hostname='hop', node_type='hop', peers_from_control_nodes=True, listener_port=6789)
control = Instance.objects.create(hostname='abc', node_type=node_type)
assert hopaddr in control.peers.all()
assert hop in control.peers.all()
hop.delete()
assert not control.peers.exists()
@@ -440,50 +203,26 @@ class TestPeers:
if a new node comes online, other peer relationships should
remain intact
"""
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', canonical=True)
hop1.peers.add(hop2addr)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop', listener_port=6789, peers_from_control_nodes=True)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop', listener_port=6789, peers_from_control_nodes=False)
hop1.peers.add(hop2)
# a control node is added
Instance.objects.create(hostname='control', node_type=node_type)
Instance.objects.create(hostname='control', node_type=node_type, listener_port=None)
assert hop1.peers.exists()
def test_reverse_peers(self, admin_user, get):
"""
if hop1 peers to hop2, hop1 should
be in hop2's reverse_peers list
"""
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', canonical=True)
hop1.peers.add(hop2addr)
resp = get(
url=reverse('api:instance_detail', kwargs={'pk': hop2.pk}),
user=admin_user,
expect=200,
)
assert hop1.pk in resp.data['reverse_peers']
def test_group_vars(self):
def test_group_vars(self, get, admin_user):
"""
control > hop1 > hop2 < execution
"""
control = Instance.objects.create(hostname='control', node_type='control')
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
ReceptorAddress.objects.create(instance=hop1, address='hop1', peers_from_control_nodes=True, port=6789, canonical=True)
control = Instance.objects.create(hostname='control', node_type='control', listener_port=None)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop', listener_port=6789, peers_from_control_nodes=True)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop', listener_port=6789, peers_from_control_nodes=False)
execution = Instance.objects.create(hostname='execution', node_type='execution', listener_port=6789)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', peers_from_control_nodes=False, port=6789, canonical=True)
execution = Instance.objects.create(hostname='execution', node_type='execution')
ReceptorAddress.objects.create(instance=execution, address='execution', peers_from_control_nodes=False, port=6789, canonical=True)
execution.peers.add(hop2addr)
hop1.peers.add(hop2addr)
execution.peers.add(hop2)
hop1.peers.add(hop2)
control_vars = yaml.safe_load(generate_group_vars_all_yml(control))
hop1_vars = yaml.safe_load(generate_group_vars_all_yml(hop1))
@@ -526,15 +265,13 @@ class TestPeers:
control = Instance.objects.create(hostname='control1', node_type='control')
write_method.assert_not_called()
# new address with peers_from_control_nodes False (no)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop1addr = ReceptorAddress.objects.create(instance=hop1, address='hop1', peers_from_control_nodes=False, canonical=True)
# new hop node with peers_from_control_nodes False (no)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop', listener_port=6789, peers_from_control_nodes=False)
hop1.delete()
write_method.assert_not_called()
# new address with peers_from_control_nodes True (yes)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop')
hop1addr = ReceptorAddress.objects.create(instance=hop1, address='hop1', peers_from_control_nodes=True, canonical=True)
# new hop node with peers_from_control_nodes True (yes)
hop1 = Instance.objects.create(hostname='hop1', node_type='hop', listener_port=6789, peers_from_control_nodes=True)
write_method.assert_called()
write_method.reset_mock()
@@ -543,21 +280,20 @@ class TestPeers:
write_method.assert_called()
write_method.reset_mock()
# new address with peers_from_control_nodes False and peered to another hop node (no)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop')
ReceptorAddress.objects.create(instance=hop2, address='hop2', peers_from_control_nodes=False, canonical=True)
hop2.peers.add(hop1addr)
# new hop node with peers_from_control_nodes False and peered to another hop node (no)
hop2 = Instance.objects.create(hostname='hop2', node_type='hop', listener_port=6789, peers_from_control_nodes=False)
hop2.peers.add(hop1)
hop2.delete()
write_method.assert_not_called()
# changing peers_from_control_nodes to False (yes)
hop1addr.peers_from_control_nodes = False
hop1addr.save()
hop1.peers_from_control_nodes = False
hop1.save()
write_method.assert_called()
write_method.reset_mock()
# deleting address that has peers_from_control_nodes to False (no)
hop1.delete() # cascade deletes to hop1addr
# deleting hop node that has peers_from_control_nodes to False (no)
hop1.delete()
write_method.assert_not_called()
# deleting control nodes (no)
@@ -579,8 +315,8 @@ class TestPeers:
# not peered, so config file should not be updated
for i in range(3):
inst = Instance.objects.create(hostname=f"exNo-{i}", node_type='execution')
ReceptorAddress.objects.create(instance=inst, address=f"exNo-{i}", port=6789, peers_from_control_nodes=False, canonical=True)
Instance.objects.create(hostname=f"exNo-{i}", node_type='execution', listener_port=6789, peers_from_control_nodes=False)
_, should_update = generate_config_data()
assert not should_update
@@ -588,13 +324,11 @@ class TestPeers:
expected_peers = []
for i in range(3):
expected_peers.append(f"hop-{i}:6789")
inst = Instance.objects.create(hostname=f"hop-{i}", node_type='hop')
ReceptorAddress.objects.create(instance=inst, address=f"hop-{i}", port=6789, peers_from_control_nodes=True, canonical=True)
Instance.objects.create(hostname=f"hop-{i}", node_type='hop', listener_port=6789, peers_from_control_nodes=True)
for i in range(3):
expected_peers.append(f"exYes-{i}:6789")
inst = Instance.objects.create(hostname=f"exYes-{i}", node_type='execution')
ReceptorAddress.objects.create(instance=inst, address=f"exYes-{i}", port=6789, peers_from_control_nodes=True, canonical=True)
Instance.objects.create(hostname=f"exYes-{i}", node_type='execution', listener_port=6789, peers_from_control_nodes=True)
new_config, should_update = generate_config_data()
assert should_update

View File

@@ -101,7 +101,6 @@ def test_default_cred_types():
'satellite6',
'scm',
'ssh',
'terraform',
'thycotic_dsv',
'thycotic_tss',
'vault',

View File

@@ -1,30 +0,0 @@
import pytest
from awx.main.models import Instance, ReceptorAddress, InstanceLink
from awx.main.tasks.system import inspect_established_receptor_connections
@pytest.mark.django_db
class TestLinkState:
@pytest.fixture(autouse=True)
def configure_settings(self, settings):
settings.IS_K8S = True
def test_inspect_established_receptor_connections(self):
'''
Change link state from ADDING to ESTABLISHED
if the receptor status KnownConnectionCosts field
has an entry for the source and target node.
'''
hop1 = Instance.objects.create(hostname='hop1')
hop2 = Instance.objects.create(hostname='hop2')
hop2addr = ReceptorAddress.objects.create(instance=hop2, address='hop2', port=5678)
InstanceLink.objects.create(source=hop1, target=hop2addr, link_state=InstanceLink.States.ADDING)
# calling with empty KnownConnectionCosts should not change the link state
inspect_established_receptor_connections({"KnownConnectionCosts": {}})
assert InstanceLink.objects.get(source=hop1, target=hop2addr).link_state == InstanceLink.States.ADDING
mesh_state = {"KnownConnectionCosts": {"hop1": {"hop2": 1}}}
inspect_established_receptor_connections(mesh_state)
assert InstanceLink.objects.get(source=hop1, target=hop2addr).link_state == InstanceLink.States.ESTABLISHED

View File

@@ -42,29 +42,3 @@ class TestMigrationSmoke:
final_state = migrator.apply_tested_migration(final_migration)
Instance = final_state.apps.get_model('main', 'Instance')
assert Instance.objects.filter(hostname='foobar').count() == 1
def test_receptor_address(self, migrator):
old_state = migrator.apply_initial_migration(('main', '0188_add_bitbucket_dc_webhook'))
Instance = old_state.apps.get_model('main', 'Instance')
for i in range(3):
Instance.objects.create(hostname=f'foobar{i}', node_type='hop')
foo = Instance.objects.create(hostname='foo', node_type='execution', listener_port=1234)
bar = Instance.objects.create(hostname='bar', node_type='execution', listener_port=None)
bar.peers.add(foo)
new_state = migrator.apply_tested_migration(
('main', '0189_inbound_hop_nodes'),
)
Instance = new_state.apps.get_model('main', 'Instance')
ReceptorAddress = new_state.apps.get_model('main', 'ReceptorAddress')
# We can now test how our migration worked, new field is there:
assert ReceptorAddress.objects.filter(address='foo', port=1234).count() == 1
assert not ReceptorAddress.objects.filter(address='bar').exists()
bar = Instance.objects.get(hostname='bar')
fooaddr = ReceptorAddress.objects.get(address='foo')
bar_peers = bar.peers.all()
assert len(bar_peers) == 1
assert fooaddr in bar_peers

View File

@@ -1,32 +0,0 @@
from awx.main.models import ReceptorAddress
import pytest
ReceptorAddress()
@pytest.mark.parametrize(
'address, protocol, port, websocket_path, expected',
[
('foo', 'tcp', 27199, '', 'foo:27199'),
('bar', 'ws', 6789, '', 'wss://bar:6789'),
('mal', 'ws', 6789, 'path', 'wss://mal:6789/path'),
('example.com', 'ws', 443, 'path', 'wss://example.com:443/path'),
],
)
def test_get_full_address(address, protocol, port, websocket_path, expected):
receptor_address = ReceptorAddress(address=address, protocol=protocol, port=port, websocket_path=websocket_path)
assert receptor_address.get_full_address() == expected
@pytest.mark.parametrize(
'protocol, expected',
[
('tcp', 'tcp-peer'),
('ws', 'ws-peer'),
('wss', 'ws-peer'),
('foo', None),
],
)
def test_get_peer_type(protocol, expected):
receptor_address = ReceptorAddress(protocol=protocol)
assert receptor_address.get_peer_type() == expected

View File

@@ -1085,27 +1085,6 @@ class TestJobCredentials(TestJobExecution):
assert open(env['ANSIBLE_NET_SSH_KEYFILE'], 'r').read() == self.EXAMPLE_PRIVATE_KEY
assert safe_env['ANSIBLE_NET_PASSWORD'] == HIDDEN_PASSWORD
def test_terraform_cloud_credentials(self, job, private_data_dir, mock_me):
terraform = CredentialType.defaults['terraform']()
hcl_config = '''
backend "s3" {
bucket = "s3_sample_bucket"
key = "/tf_state/"
region = "us-east-1"
}
'''
credential = Credential(pk=1, credential_type=terraform, inputs={'configuration': hcl_config})
credential.inputs['configuration'] = encrypt_field(credential, 'configuration')
job.credentials.add(credential)
env = {}
safe_env = {}
credential.credential_type.inject_credential(credential, env, safe_env, [], private_data_dir)
local_path = to_host_path(env['TF_BACKEND_CONFIG_FILE'], private_data_dir)
config = open(local_path, 'r').read()
assert config == hcl_config
def test_custom_environment_injectors_with_jinja_syntax_error(self, private_data_dir, mock_me):
some_cloud = CredentialType(
kind='cloud',

View File

@@ -20,6 +20,7 @@ from awx.main.analytics.broadcast_websocket import (
RelayWebsocketStats,
RelayWebsocketStatsManager,
)
import awx.main.analytics.subsystem_metrics as s_metrics
logger = logging.getLogger('awx.main.wsrelay')
@@ -53,6 +54,7 @@ class WebsocketRelayConnection:
self.protocol = protocol
self.verify_ssl = verify_ssl
self.channel_layer = None
self.subsystem_metrics = s_metrics.Metrics(instance_name=name)
self.producers = dict()
self.connected = False
@@ -339,7 +341,7 @@ class WebSocketRelayManager(object):
if deleted_remote_hosts:
logger.info(f"Removing {deleted_remote_hosts} from websocket broadcast list")
await asyncio.gather(*[self.cleanup_offline_host(h) for h in deleted_remote_hosts])
await asyncio.gather(self.cleanup_offline_host(h) for h in deleted_remote_hosts)
if new_remote_hosts:
logger.info(f"Adding {new_remote_hosts} to websocket broadcast list")

View File

@@ -1076,35 +1076,6 @@ HOST_METRIC_SUMMARY_TASK_LAST_TS = None
HOST_METRIC_SUMMARY_TASK_INTERVAL = 7 # days
# TODO: cmeyers, replace with with register pattern
# The register pattern is particularly nice for this because we need
# to know the process to start the thread that will be the server.
# The registration location should be the same location as we would
# call MetricsServer.start()
# Note: if we don't get to this TODO, then at least create constants
# for the services strings below.
# TODO: cmeyers, break this out into a separate django app so other
# projects can take advantage.
METRICS_SERVICE_CALLBACK_RECEIVER = 'callback_receiver'
METRICS_SERVICE_DISPATCHER = 'dispatcher'
METRICS_SERVICE_WEBSOCKETS = 'websockets'
METRICS_SUBSYSTEM_CONFIG = {
'server': {
METRICS_SERVICE_CALLBACK_RECEIVER: {
'port': 8014,
},
METRICS_SERVICE_DISPATCHER: {
'port': 8015,
},
METRICS_SERVICE_WEBSOCKETS: {
'port': 8016,
},
}
}
# django-ansible-base
ANSIBLE_BASE_TEAM_MODEL = 'main.Team'
ANSIBLE_BASE_ORGANIZATION_MODEL = 'main.Organization'

View File

@@ -21,7 +21,7 @@ from split_settings.tools import optional, include
from .defaults import * # NOQA
# awx-manage shell_plus --notebook
NOTEBOOK_ARGUMENTS = ['--NotebookApp.token=', '--ip', '0.0.0.0', '--port', '9888', '--allow-root', '--no-browser']
NOTEBOOK_ARGUMENTS = ['--NotebookApp.token=', '--ip', '0.0.0.0', '--port', '8888', '--allow-root', '--no-browser']
# print SQL queries in shell_plus
SHELL_PLUS_PRINT_SQL = False

View File

@@ -29,7 +29,6 @@ import Notifications from './models/Notifications';
import Organizations from './models/Organizations';
import ProjectUpdates from './models/ProjectUpdates';
import Projects from './models/Projects';
import ReceptorAddresses from './models/Receptor';
import Roles from './models/Roles';
import Root from './models/Root';
import Schedules from './models/Schedules';
@@ -80,7 +79,6 @@ const NotificationsAPI = new Notifications();
const OrganizationsAPI = new Organizations();
const ProjectUpdatesAPI = new ProjectUpdates();
const ProjectsAPI = new Projects();
const ReceptorAPI = new ReceptorAddresses();
const RolesAPI = new Roles();
const RootAPI = new Root();
const SchedulesAPI = new Schedules();
@@ -132,7 +130,6 @@ export {
OrganizationsAPI,
ProjectUpdatesAPI,
ProjectsAPI,
ReceptorAPI,
RolesAPI,
RootAPI,
SchedulesAPI,

View File

@@ -8,7 +8,6 @@ class Instances extends Base {
this.readHealthCheckDetail = this.readHealthCheckDetail.bind(this);
this.healthCheck = this.healthCheck.bind(this);
this.readInstanceGroup = this.readInstanceGroup.bind(this);
this.readReceptorAddresses = this.readReceptorAddresses.bind(this);
this.deprovisionInstance = this.deprovisionInstance.bind(this);
}
@@ -28,17 +27,6 @@ class Instances extends Base {
return this.http.get(`${this.baseUrl}${instanceId}/instance_groups/`);
}
readReceptorAddresses(instanceId) {
return this.http.get(`${this.baseUrl}${instanceId}/receptor_addresses/`);
}
updateReceptorAddresses(instanceId, data) {
return this.http.post(
`${this.baseUrl}${instanceId}/receptor_addresses/`,
data
);
}
deprovisionInstance(instanceId) {
return this.http.patch(`${this.baseUrl}${instanceId}/`, {
node_state: 'deprovisioning',

View File

@@ -1,14 +0,0 @@
import Base from '../Base';
class ReceptorAddresses extends Base {
constructor(http) {
super(http);
this.baseUrl = 'api/v2/receptor_addresses/';
}
updateReceptorAddresses(instanceId, data) {
return this.http.post(`${this.baseUrl}`, data);
}
}
export default ReceptorAddresses;

View File

@@ -12,7 +12,6 @@ import { SettingsAPI } from 'api';
import ContentLoading from 'components/ContentLoading';
import InstanceDetail from './InstanceDetail';
import InstancePeerList from './InstancePeers';
import InstanceListenerAddressList from './InstanceListenerAddressList';
function Instance({ setBreadcrumb }) {
const { me } = useConfig();
@@ -55,12 +54,7 @@ function Instance({ setBreadcrumb }) {
}, [request]);
if (isK8s) {
tabsArray.push({
name: t`Listener Addresses`,
link: `${match.url}/listener_addresses`,
id: 1,
});
tabsArray.push({ name: t`Peers`, link: `${match.url}/peers`, id: 2 });
tabsArray.push({ name: t`Peers`, link: `${match.url}/peers`, id: 1 });
}
if (isLoading) {
return <ContentLoading />;
@@ -78,14 +72,6 @@ function Instance({ setBreadcrumb }) {
<Route path="/instances/:id/details" key="details">
<InstanceDetail isK8s={isK8s} setBreadcrumb={setBreadcrumb} />
</Route>
{isK8s && (
<Route
path="/instances/:id/listener_addresses"
key="listener_addresses"
>
<InstanceListenerAddressList setBreadcrumb={setBreadcrumb} />
</Route>
)}
{isK8s && (
<Route path="/instances/:id/peers" key="peers">
<InstancePeerList setBreadcrumb={setBreadcrumb} />

View File

@@ -9,10 +9,6 @@ function InstanceAdd() {
const [formError, setFormError] = useState();
const handleSubmit = async (values) => {
try {
if (values.listener_port === undefined) {
values.listener_port = null;
}
const {
data: { id },
} = await InstancesAPI.create(values);

View File

@@ -36,7 +36,6 @@ describe('<InstanceAdd />', () => {
});
});
expect(InstancesAPI.create).toHaveBeenCalledWith({
listener_port: null, // injected if listener_port is not set
node_type: 'hop',
});
expect(history.location.pathname).toBe('/instances/13/details');

View File

@@ -183,7 +183,6 @@ function InstanceDetail({ setBreadcrumb, isK8s }) {
}
const isHopNode = instance.node_type === 'hop';
const isExecutionNode = instance.node_type === 'execution';
const isManaged = instance.managed;
return (
<>
@@ -209,31 +208,33 @@ function InstanceDetail({ setBreadcrumb, isK8s }) {
<Detail label={t`Node Type`} value={instance.node_type} />
<Detail label={t`Host`} value={instance.ip_address} />
<Detail label={t`Listener Port`} value={instance.listener_port} />
{!isManaged && instance.related?.install_bundle && (
<Detail
label={t`Install Bundle`}
value={
<Tooltip content={t`Click to download bundle`}>
<Button
component="a"
isSmall
href={`${instance.related?.install_bundle}`}
target="_blank"
variant="secondary"
dataCy="install-bundle-download-button"
rel="noopener noreferrer"
>
<DownloadIcon />
</Button>
</Tooltip>
}
/>
)}
{(isExecutionNode || isHopNode) && (
<Detail
label={t`Peers from control nodes`}
value={instance.peers_from_control_nodes ? t`On` : t`Off`}
/>
<>
{instance.related?.install_bundle && (
<Detail
label={t`Install Bundle`}
value={
<Tooltip content={t`Click to download bundle`}>
<Button
component="a"
isSmall
href={`${instance.related?.install_bundle}`}
target="_blank"
variant="secondary"
dataCy="install-bundle-download-button"
rel="noopener noreferrer"
>
<DownloadIcon />
</Button>
</Tooltip>
}
/>
)}
<Detail
label={t`Peers from control nodes`}
value={instance.peers_from_control_nodes ? t`On` : t`Off`}
/>
</>
)}
{!isHopNode && (
<>
@@ -293,9 +294,7 @@ function InstanceDetail({ setBreadcrumb, isK8s }) {
value={instance.capacity_adjustment}
onChange={handleChangeValue}
isDisabled={
!config?.me?.is_superuser ||
!instance.enabled ||
!isManaged
!config?.me?.is_superuser || !instance.enabled
}
data-cy="slider"
/>
@@ -339,31 +338,31 @@ function InstanceDetail({ setBreadcrumb, isK8s }) {
)}
</DetailList>
<CardActionsRow>
{config?.me?.is_superuser && isK8s && !isManaged && (
<>
<Button
ouiaId="instance-detail-edit-button"
aria-label={t`edit`}
component={Link}
to={`/instances/${id}/edit`}
>
{t`Edit`}
</Button>
{config?.me?.is_superuser && isK8s && (isExecutionNode || isHopNode) && (
<Button
ouiaId="instance-detail-edit-button"
aria-label={t`edit`}
component={Link}
to={`/instances/${id}/edit`}
>
{t`Edit`}
</Button>
)}
{config?.me?.is_superuser &&
isK8s &&
(isExecutionNode || isHopNode) && (
<RemoveInstanceButton
dataCy="remove-instance-button"
itemsToRemove={[instance]}
isK8s={isK8s}
onRemove={removeInstances}
/>
</>
)}
)}
{isExecutionNode && (
<Tooltip content={t`Run a health check on the instance`}>
<Button
isDisabled={
!config?.me?.is_superuser ||
instance.health_check_pending ||
instance.managed
!config?.me?.is_superuser || instance.health_check_pending
}
variant="primary"
ouiaId="health-check-button"
@@ -377,14 +376,12 @@ function InstanceDetail({ setBreadcrumb, isK8s }) {
</Button>
</Tooltip>
)}
{!isHopNode && (
<InstanceToggle
css="display: inline-flex;"
fetchInstances={fetchDetails}
instance={instance}
dataCy="enable-instance"
/>
)}
<InstanceToggle
css="display: inline-flex;"
fetchInstances={fetchDetails}
instance={instance}
dataCy="enable-instance"
/>
</CardActionsRow>
{error && (

View File

@@ -48,7 +48,6 @@ describe('<InstanceDetail/>', () => {
cpu_capacity: 32,
mem_capacity: 38,
enabled: true,
managed: false,
managed_by_policy: true,
node_type: 'execution',
node_state: 'ready',

View File

@@ -114,8 +114,7 @@ function InstanceListItem({
);
const isHopNode = instance.node_type === 'hop';
const isManaged = instance.managed;
const isExecutionNode = instance.node_type === 'execution';
return (
<>
<Tr
@@ -139,7 +138,7 @@ function InstanceListItem({
rowIndex,
isSelected,
onSelect,
disable: isManaged,
disable: !(isExecutionNode || isHopNode),
}}
dataLabel={t`Selected`}
/>

View File

@@ -1,154 +0,0 @@
import React, { useCallback, useEffect } from 'react';
import { t } from '@lingui/macro';
import { CardBody } from 'components/Card';
import PaginatedTable, {
getSearchableKeys,
HeaderCell,
HeaderRow,
// ToolbarAddButton,
} from 'components/PaginatedTable';
import useToast from 'hooks/useToast';
import { getQSConfig } from 'util/qs';
import { useParams } from 'react-router-dom';
import useRequest from 'hooks/useRequest';
import DataListToolbar from 'components/DataListToolbar';
import { InstancesAPI, ReceptorAPI } from 'api';
import useSelected from 'hooks/useSelected';
import InstanceListenerAddressListItem from './InstanceListenerAddressListItem';
const QS_CONFIG = getQSConfig('peer', {
page: 1,
page_size: 20,
order_by: 'pk',
});
function InstanceListenerAddressList({ setBreadcrumb }) {
const { id } = useParams();
const { Toast, toastProps } = useToast();
const {
isLoading,
error: contentError,
request: fetchListenerAddresses,
result: {
instance,
listenerAddresses,
count,
relatedSearchableKeys,
searchableKeys,
},
} = useRequest(
useCallback(async () => {
const [
{ data: detail },
{
data: { results },
},
actions,
] = await Promise.all([
InstancesAPI.readDetail(id),
ReceptorAPI.read(),
InstancesAPI.readOptions(),
]);
const listenerAddress_list = [];
for (let q = 0; q < results.length; q++) {
const receptor = results[q];
if (receptor.managed === true) continue;
if (id.toString() === receptor.instance.toString()) {
receptor.name = detail.hostname;
listenerAddress_list.push(receptor);
}
}
return {
instance: detail,
listenerAddresses: listenerAddress_list,
count: listenerAddress_list.length,
relatedSearchableKeys: (actions?.data?.related_search_fields || []).map(
(val) => val.slice(0, -8)
),
searchableKeys: getSearchableKeys(actions.data.actions?.GET),
};
}, [id]),
{
instance: {},
listenerAddresses: [],
count: 0,
relatedSearchableKeys: [],
searchableKeys: [],
}
);
useEffect(() => {
fetchListenerAddresses();
}, [fetchListenerAddresses]);
useEffect(() => {
if (instance) {
setBreadcrumb(instance);
}
}, [instance, setBreadcrumb]);
const { selected, isAllSelected, handleSelect, clearSelected, selectAll } =
useSelected(listenerAddresses);
return (
<CardBody>
<PaginatedTable
contentError={contentError}
hasContentLoading={isLoading}
items={listenerAddresses}
itemCount={count}
pluralizedItemName={t`Listener Addresses`}
qsConfig={QS_CONFIG}
onRowClick={handleSelect}
clearSelected={clearSelected}
toolbarSearchableKeys={searchableKeys}
toolbarRelatedSearchableKeys={relatedSearchableKeys}
toolbarSearchColumns={[
{
name: t`Name`,
key: 'hostname__icontains',
isDefault: true,
},
]}
toolbarSortColumns={[
{
name: t`Name`,
key: 'hostname',
},
]}
headerRow={
<HeaderRow qsConfig={QS_CONFIG}>
<HeaderCell sortKey="address">{t`Address`}</HeaderCell>
<HeaderCell sortKey="port">{t`Port`}</HeaderCell>
<HeaderCell sortKey="protocol">{t`Protocol`}</HeaderCell>
<HeaderCell sortKey="canonical">{t`Canonical`}</HeaderCell>
</HeaderRow>
}
renderToolbar={(props) => (
<DataListToolbar
{...props}
isAllSelected={isAllSelected}
onSelectAll={selectAll}
qsConfig={QS_CONFIG}
additionalControls={[]}
/>
)}
renderRow={(listenerAddress, index) => (
<InstanceListenerAddressListItem
isSelected={selected.some((row) => row.id === listenerAddress.id)}
onSelect={() => handleSelect(listenerAddress)}
key={listenerAddress.id}
peerListenerAddress={listenerAddress}
rowIndex={index}
/>
)}
/>
<Toast {...toastProps} />
</CardBody>
);
}
export default InstanceListenerAddressList;

View File

@@ -1,46 +0,0 @@
import React from 'react';
import { t } from '@lingui/macro';
import 'styled-components/macro';
import { Tr, Td } from '@patternfly/react-table';
function InstanceListenerAddressListItem({
peerListenerAddress,
isSelected,
onSelect,
rowIndex,
}) {
const labelId = `check-action-${peerListenerAddress.id}`;
return (
<Tr
id={`peerListenerAddress-row-${peerListenerAddress.id}`}
ouiaId={`peerListenerAddress-row-${peerListenerAddress.id}`}
>
<Td
select={{
rowIndex,
isSelected,
onSelect,
}}
dataLabel={t`Selected`}
/>
<Td id={labelId} dataLabel={t`Address`}>
{peerListenerAddress.address}
</Td>
<Td id={labelId} dataLabel={t`Port`}>
{peerListenerAddress.port}
</Td>
<Td id={labelId} dataLabel={t`Protocol`}>
{peerListenerAddress.protocol}
</Td>
<Td id={labelId} dataLabel={t`Canonical`}>
{peerListenerAddress.canonical.toString()}
</Td>
</Tr>
);
}
export default InstanceListenerAddressListItem;

View File

@@ -1 +0,0 @@
export { default } from './InstanceListenerAddressList';

View File

@@ -16,7 +16,7 @@ import { getQSConfig, parseQueryString, mergeParams } from 'util/qs';
import { useLocation, useParams } from 'react-router-dom';
import useRequest, { useDismissableError } from 'hooks/useRequest';
import DataListToolbar from 'components/DataListToolbar';
import { InstancesAPI, ReceptorAPI } from 'api';
import { InstancesAPI } from 'api';
import useExpanded from 'hooks/useExpanded';
import useSelected from 'hooks/useSelected';
import InstancePeerListItem from './InstancePeerListItem';
@@ -24,7 +24,7 @@ import InstancePeerListItem from './InstancePeerListItem';
const QS_CONFIG = getQSConfig('peer', {
page: 1,
page_size: 20,
order_by: 'pk',
order_by: 'hostname',
});
function InstancePeerList({ setBreadcrumb }) {
@@ -47,35 +47,18 @@ function InstancePeerList({ setBreadcrumb }) {
const [
{ data: detail },
{
data: { results },
data: { results, count: itemNumber },
},
actions,
instances,
] = await Promise.all([
InstancesAPI.readDetail(id),
InstancesAPI.readPeers(id, params),
InstancesAPI.readOptions(),
InstancesAPI.read(),
]);
const address_list = [];
for (let q = 0; q < results.length; q++) {
const receptor = results[q];
if (receptor.managed === true) continue;
const host = instances.data.results.filter(
(obj) => obj.id === receptor.instance
)[0];
const copy = receptor;
copy.hostname = host.hostname;
copy.node_type = host.node_type;
address_list.push(copy);
}
return {
instance: detail,
peers: address_list,
count: address_list.length,
peers: results,
count: itemNumber,
relatedSearchableKeys: (actions?.data?.related_search_fields || []).map(
(val) => val.slice(0, -8)
),
@@ -107,73 +90,15 @@ function InstancePeerList({ setBreadcrumb }) {
useSelected(peers);
const fetchInstancesToAssociate = useCallback(
async (params) => {
const address_list = [];
const instances = await InstancesAPI.read(
(params) =>
InstancesAPI.read(
mergeParams(params, {
...{ not__id: id },
...{ not__node_type: ['control', 'hybrid'] },
...{ not__hostname: instance.peers },
})
);
const receptors = (await ReceptorAPI.read()).data.results;
// get instance ids of the current peered receptor ids
const already_peered_instance_ids = [];
for (let h = 0; h < instance.peers.length; h++) {
const matched = receptors.filter((obj) => obj.id === instance.peers[h]);
matched.forEach((element) => {
already_peered_instance_ids.push(element.instance);
});
}
for (let q = 0; q < receptors.length; q++) {
const receptor = receptors[q];
if (already_peered_instance_ids.includes(receptor.instance)) {
// ignore reverse peers
continue;
}
if (instance.peers.includes(receptor.id)) {
// no links to existing links
continue;
}
if (instance.id === receptor.instance) {
// no links to thy self
continue;
}
if (instance.managed) {
// no managed nodes
continue;
}
const host = instances.data.results.filter(
(obj) => obj.id === receptor.instance
)[0];
if (host === undefined) {
// no hosts
continue;
}
if (receptor.is_internal) {
continue;
}
const copy = receptor;
copy.hostname = host.hostname;
copy.node_type = host.node_type;
copy.canonical = copy.canonical.toString();
address_list.push(copy);
}
instances.data.results = address_list;
return instances;
},
[instance]
),
[id, instance]
);
const {
@@ -183,15 +108,17 @@ function InstancePeerList({ setBreadcrumb }) {
} = useRequest(
useCallback(
async (instancesPeerToAssociate) => {
const selected_peers = instancesPeerToAssociate.map((obj) => obj.id);
const new_peers = [...new Set([...instance.peers, ...selected_peers])];
const selected_hostname = instancesPeerToAssociate.map(
(obj) => obj.hostname
);
const new_peers = [
...new Set([...instance.peers, ...selected_hostname]),
];
await InstancesAPI.update(instance.id, { peers: new_peers });
fetchPeers();
addToast({
id: instancesPeerToAssociate,
title: t`Peers update on ${instance.hostname}. Please be sure to run the install bundle for ${instance.hostname} again in order to see changes take effect.`,
title: t`${selected_hostname} added as a peer. Please be sure to run the install bundle for ${instance.hostname} again in order to see changes take effect.`,
variant: AlertVariant.success,
hasTimeout: true,
});
@@ -206,18 +133,17 @@ function InstancePeerList({ setBreadcrumb }) {
error: disassociateError,
} = useRequest(
useCallback(async () => {
let new_peers = instance.peers;
const selected_ids = selected.map((obj) => obj.id);
for (let i = 0; i < selected_ids.length; i++) {
new_peers = new_peers.filter((s_id) => s_id !== selected_ids[i]);
const new_peers = [];
const selected_hostname = selected.map((obj) => obj.hostname);
for (let i = 0; i < instance.peers.length; i++) {
if (!selected_hostname.includes(instance.peers[i])) {
new_peers.push(instance.peers[i]);
}
}
await InstancesAPI.update(instance.id, { peers: new_peers });
fetchPeers();
addToast({
title: t`Peer removed. Please be sure to run the install bundle for ${instance.hostname} again in order to see changes take effect.`,
title: t`${selected_hostname} removed. Please be sure to run the install bundle for ${instance.hostname} again in order to see changes take effect.`,
variant: AlertVariant.success,
hasTimeout: true,
});
@@ -264,11 +190,9 @@ function InstancePeerList({ setBreadcrumb }) {
<HeaderCell
tooltip={t`Cannot run health check on hop nodes.`}
sortKey="hostname"
>{t`Instance Name`}</HeaderCell>
<HeaderCell sortKey="address">{t`Address`}</HeaderCell>
<HeaderCell sortKey="port">{t`Port`}</HeaderCell>
>{t`Name`}</HeaderCell>
<HeaderCell sortKey="errors">{t`Status`}</HeaderCell>
<HeaderCell sortKey="node_type">{t`Node Type`}</HeaderCell>
<HeaderCell sortKey="canonical">{t`Canonical`}</HeaderCell>
</HeaderRow>
}
renderToolbar={(props) => (
@@ -294,7 +218,7 @@ function InstancePeerList({ setBreadcrumb }) {
key="disassociate"
onDisassociate={handlePeersDiassociate}
itemsToDisassociate={selected}
modalTitle={t`Remove peers?`}
modalTitle={t`Remove instance from peers?`}
/>
),
]}
@@ -319,15 +243,12 @@ function InstancePeerList({ setBreadcrumb }) {
isModalOpen={isModalOpen}
onAssociate={handlePeerAssociate}
onClose={() => setIsModalOpen(false)}
title={t`Select Peer Addresses`}
title={t`Select Instances`}
optionsRequest={readInstancesOptions}
displayKey="hostname"
columns={[
{ key: 'hostname', name: t`Name` },
{ key: 'address', name: t`Address` },
{ key: 'port', name: t`Port` },
{ key: 'node_type', name: t`Node Type` },
{ key: 'protocol', name: t`Protocol` },
]}
/>
)}

View File

@@ -2,8 +2,10 @@ import React from 'react';
import { Link } from 'react-router-dom';
import { t } from '@lingui/macro';
import 'styled-components/macro';
import { Tooltip } from '@patternfly/react-core';
import { Tr, Td, ExpandableRowContent } from '@patternfly/react-table';
import { formatDateString } from 'util/dates';
import StatusLabel from 'components/StatusLabel';
import { Detail, DetailList } from 'components/DetailList';
function InstancePeerListItem({
@@ -41,26 +43,29 @@ function InstancePeerListItem({
}}
dataLabel={t`Selected`}
/>
<Td id={labelId} dataLabel={t`Name`}>
<Link to={`/instances/${peerInstance.instance}/details`}>
<Link to={`/instances/${peerInstance.id}/details`}>
<b>{peerInstance.hostname}</b>
</Link>
</Td>
<Td id={labelId} dataLabel={t`Address`}>
{peerInstance.address}
</Td>
<Td id={labelId} dataLabel={t`Port`}>
{peerInstance.port}
<Td dataLabel={t`Status`}>
<Tooltip
content={
<div>
{t`Last Health Check`}
&nbsp;
{formatDateString(
peerInstance.last_health_check ?? peerInstance.last_seen
)}
</div>
}
>
<StatusLabel status={peerInstance.node_state} />
</Tooltip>
</Td>
<Td dataLabel={t`Node Type`}>{peerInstance.node_type}</Td>
<Td id={labelId} dataLabel={t`Canonical`}>
{peerInstance.canonical.toString()}
</Td>
</Tr>
{!isHopNode && (
<Tr

View File

@@ -25,7 +25,6 @@ function Instances() {
[`/instances/${instance.id}`]: `${instance.hostname}`,
[`/instances/${instance.id}/details`]: t`Details`,
[`/instances/${instance.id}/peers`]: t`Peers`,
[`/instances/${instance.id}/listener_addresses`]: t`Listener Addresses`,
[`/instances/${instance.id}/edit`]: t`Edit Instance`,
});
}, []);

View File

@@ -1,6 +1,6 @@
import React from 'react';
import React, { useCallback } from 'react';
import { t } from '@lingui/macro';
import { Formik, useField } from 'formik';
import { Formik, useField, useFormikContext } from 'formik';
import { Form, FormGroup, CardBody } from '@patternfly/react-core';
import { FormColumnLayout } from 'components/FormLayout';
import FormField, {
@@ -9,6 +9,7 @@ import FormField, {
} from 'components/FormField';
import FormActionGroup from 'components/FormActionGroup';
import AnsibleSelect from 'components/AnsibleSelect';
import { PeersLookup } from 'components/Lookup';
import { required } from 'util/validators';
const INSTANCE_TYPES = [
@@ -22,6 +23,16 @@ function InstanceFormFields({ isEdit }) {
validate: required(t`Set a value for this field`),
});
const { setFieldValue } = useFormikContext();
const [peersField, peersMeta, peersHelpers] = useField('peers');
const handlePeersUpdate = useCallback(
(value) => {
setFieldValue('peers', value);
},
[setFieldValue]
);
return (
<>
<FormField
@@ -80,6 +91,20 @@ function InstanceFormFields({ isEdit }) {
isDisabled={isEdit}
/>
</FormGroup>
<PeersLookup
helperTextInvalid={peersMeta.error}
isValid={!peersMeta.touched || !peersMeta.error}
onBlur={() => peersHelpers.setTouched()}
onChange={handlePeersUpdate}
value={peersField.value}
tooltip={t`Select the Peers Instances.`}
fieldName="peers"
formLabel={t`Peers`}
multiple
typePeers
id="peers"
isRequired
/>
<FormGroup fieldId="instance-option-checkboxes" label={t`Options`}>
<CheckboxField
id="enabled"

View File

@@ -67,8 +67,6 @@ options:
description:
- List of peers to connect outbound to. Only configurable for hop and execution nodes.
- To remove all current peers, set value to an empty list, [].
- Each item is an ID or address of a receptor address.
- If item is address, it must be unique across all receptor addresses.
required: False
type: list
elements: str
@@ -85,24 +83,12 @@ EXAMPLES = '''
awx.awx.instance:
hostname: my-instance.prod.example.com
capacity_adjustment: 0.4
listener_port: 31337
- name: Deprovision the instance
awx.awx.instance:
hostname: my-instance.prod.example.com
node_state: deprovisioning
- name: Create execution node
awx.awx.instance:
hostname: execution.example.com
node_type: execution
peers:
- 12
- route.to.hop.example.com
- name: Remove peers
awx.awx.instance:
hostname: execution.example.com
peers:
'''
from ..module_utils.controller_api import ControllerAPIModule
@@ -138,17 +124,6 @@ def main():
# Attempt to look up an existing item based on the provided data
existing_item = module.get_one('instances', name_or_id=hostname)
# peer item can be an id or address
# if address, get the id
peers_ids = []
if peers:
for p in peers:
if not p.isdigit():
p_id = module.get_one('receptor_addresses', allow_none=False, data={'address': p})
peers_ids.append(p_id['id'])
else:
peers_ids.append(p)
# Create the data that gets sent for create and update
new_fields = {'hostname': hostname}
if capacity_adjustment is not None:
@@ -164,7 +139,7 @@ def main():
if listener_port is not None:
new_fields['listener_port'] = listener_port
if peers is not None:
new_fields['peers'] = peers_ids
new_fields['peers'] = peers
if peers_from_control_nodes is not None:
new_fields['peers_from_control_nodes'] = peers_from_control_nodes

View File

@@ -54,7 +54,7 @@ extra_endpoints = {
}
# Global module parameters we can ignore
ignore_parameters = ['state', 'new_name', 'update_secrets', 'copy_from', 'is_internal']
ignore_parameters = ['state', 'new_name', 'update_secrets', 'copy_from']
# Some modules take additional parameters that do not appear in the API
# Add the module name as the key with the value being the list of params to ignore
@@ -248,7 +248,7 @@ def test_completeness(collection_import, request, admin_user, job_template, exec
singular_endpoint = '{0}'.format(endpoint)
if singular_endpoint.endswith('ies'):
singular_endpoint = singular_endpoint[:-3]
elif singular_endpoint != 'settings' and singular_endpoint.endswith('s'):
if singular_endpoint != 'settings' and singular_endpoint.endswith('s'):
singular_endpoint = singular_endpoint[:-1]
module_name = '{0}'.format(singular_endpoint)

View File

@@ -134,17 +134,21 @@ def test_export_simple(
@pytest.mark.django_db
def test_export_system_auditor(run_module, organization, system_auditor): # noqa: F811
def test_export_system_auditor(run_module, schedule, system_auditor): # noqa: F811
"""
This test illustrates that export of resources can now happen
when ran as non-root user (i.e. system auditor). The OPTIONS
endpoint does NOT return POST for a system auditor, but now we
make a best-effort to parse the description string, which will
often have the fields.
This test illustrates that deficiency of export when ran as non-root user (i.e. system auditor).
The OPTIONS endpoint does NOT return POST for a system auditor. This is bad for the export code
because it relies on crawling the OPTIONS POST response to determine the fields to export.
"""
result = run_module('export', dict(all=True), system_auditor)
assert not result.get('failed', False), result.get('msg', result)
assert 'msg' not in result
assert 'assets' in result
assert result.get('failed', False), result.get('msg', result)
find_by(result['assets'], 'organizations', 'name', 'Default')
assert 'Failed to export assets substring not found' in result['msg'], (
'If you found this error then you have probably fixed a feature! The export code attempts to assertain the POST fields from the `description` field,'
' but both the API side and the client inference code are lacking.'
)
# r = result['assets']['schedules'][0]
# assert r['natural_key']['name'] == 'test-sched'
# assert 'rrule' not in r, 'If you found this error then you have probably fixed a feature! We WANT rrule to be found in the export schedule payload.'

View File

@@ -13,32 +13,39 @@ def test_peers_adding_and_removing(run_module, admin_user):
with override_settings(IS_K8S=True):
result = run_module(
'instance',
{'hostname': 'hopnode', 'node_type': 'hop', 'node_state': 'installed', 'listener_port': 6789},
{'hostname': 'hopnode1', 'node_type': 'hop', 'peers_from_control_nodes': True, 'node_state': 'installed', 'listener_port': 27199},
admin_user,
)
assert result['changed']
hop_node = Instance.objects.get(pk=result.get('id'))
hop_node_1 = Instance.objects.get(pk=result.get('id'))
assert hop_node.node_type == 'hop'
address = hop_node.receptor_addresses.get(pk=result.get('id'))
assert address.port == 6789
assert hop_node_1.peers_from_control_nodes is True
assert hop_node_1.node_type == 'hop'
result = run_module(
'instance',
{'hostname': 'executionnode', 'node_type': 'execution', 'node_state': 'installed', 'peers': ['hopnode']},
{'hostname': 'hopnode2', 'node_type': 'hop', 'peers_from_control_nodes': True, 'node_state': 'installed', 'listener_port': 27199},
admin_user,
)
assert result['changed']
hop_node_2 = Instance.objects.get(pk=result.get('id'))
result = run_module(
'instance',
{'hostname': 'executionnode', 'node_type': 'execution', 'node_state': 'installed', 'listener_port': 27199, 'peers': ['hopnode1', 'hopnode2']},
admin_user,
)
assert result['changed']
execution_node = Instance.objects.get(pk=result.get('id'))
assert set(execution_node.peers.all()) == {address}
assert set(execution_node.peers.all()) == {hop_node_1, hop_node_2}
result = run_module(
'instance',
{'hostname': 'executionnode', 'node_type': 'execution', 'node_state': 'installed', 'peers': []},
{'hostname': 'executionnode', 'node_type': 'execution', 'node_state': 'installed', 'listener_port': 27199, 'peers': []},
admin_user,
)

View File

@@ -155,4 +155,4 @@ def test_build_notification_message_undefined(run_module, admin_user, organizati
nt = NotificationTemplate.objects.get(id=result['id'])
body = job.build_notification_message(nt, 'running')
assert 'The template rendering return a blank body' in body[1]
assert '{"started_by": "My Placeholder"}' in body[1]

View File

@@ -42,6 +42,7 @@
node_type: execution
node_state: installed
capacity_adjustment: 0.4
listener_port: 31337
register: result
- assert:
@@ -73,9 +74,11 @@
- block:
- name: Create hop node 1
awx.awx.instance:
hostname: "{{ hostname1 }}"
hostname: hopnode1
node_type: hop
node_state: installed
listener_port: 27199
peers_from_control_nodes: True
register: result
- assert:
@@ -84,9 +87,11 @@
- name: Create hop node 2
awx.awx.instance:
hostname: "{{ hostname2 }}"
hostname: hopnode2
node_type: hop
node_state: installed
listener_port: 27199
peers_from_control_nodes: True
register: result
- assert:
@@ -95,12 +100,13 @@
- name: Create execution node
awx.awx.instance:
hostname: "{{ hostname3 }}"
hostname: executionnode
node_type: execution
node_state: installed
listener_port: 27199
peers:
- "{{ hostname1 }}"
- "{{ hostname2 }}"
- "hopnode1"
- "hopnode2"
register: result
- assert:
@@ -109,9 +115,10 @@
- name: Remove execution node peers
awx.awx.instance:
hostname: "{{ hostname3 }}"
hostname: executionnode
node_type: execution
node_state: installed
listener_port: 27199
peers: []
register: result
@@ -119,15 +126,4 @@
that:
- result is changed
always:
- name: Deprovision the instances
awx.awx.instance:
hostname: "{{ item }}"
node_state: deprovisioning
with_items:
- "{{ hostname1 }}"
- "{{ hostname2 }}"
- "{{ hostname3 }}"
when: IS_K8S

View File

@@ -43,8 +43,8 @@ class Connection(object):
self.session = requests.Session()
self.uses_session_cookie = False
def get_session_requirements(self, next=config.api_base_path):
self.get(config.api_base_path) # this causes a cookie w/ the CSRF token to be set
def get_session_requirements(self, next='/api/'):
self.get('/api/') # this causes a cookie w/ the CSRF token to be set
return dict(next=next)
def login(self, username=None, password=None, token=None, **kwargs):
@@ -52,7 +52,7 @@ class Connection(object):
_next = kwargs.get('next')
if _next:
headers = self.session.headers.copy()
response = self.post(f"{config.api_base_path}login/", headers=headers, data=dict(username=username, password=password, next=_next))
response = self.post('/api/login/', headers=headers, data=dict(username=username, password=password, next=_next))
# The login causes a redirect so we need to search the history of the request to find the header
for historical_response in response.history:
if 'X-API-Session-Cookie-Name' in historical_response.headers:

View File

@@ -3,7 +3,6 @@ import json
from awxkit.utils import poll_until
from awxkit.exceptions import WaitUntilTimeout
from awxkit.config import config
def bytes_to_str(obj):
@@ -84,7 +83,7 @@ class HasStatus(object):
if getattr(self, 'job_explanation', '').startswith('Previous Task Failed'):
try:
data = json.loads(self.job_explanation.replace('Previous Task Failed: ', ''))
dependency = self.walk('/{0}v2/{1}s/{2}/'.format(config.api_base_path, data['job_type'], data['job_id']))
dependency = self.walk('/api/v2/{0}s/{1}/'.format(data['job_type'], data['job_id']))
if hasattr(dependency, 'failure_output_details'):
msg += '\nDependency output:\n{}'.format(dependency.failure_output_details())
else:

View File

@@ -199,7 +199,7 @@ class ApiV2(base.Base):
return None
fields['natural_key'] = natural_key
return fields
return utils.remove_encrypted(fields)
def _export_list(self, endpoint):
post_fields = utils.get_post_fields(endpoint, self._cache)
@@ -280,7 +280,7 @@ class ApiV2(base.Base):
_page = self._cache.get_by_natural_key(value)
post_data[field] = _page['id'] if _page is not None else None
else:
post_data[field] = utils.remove_encrypted(value)
post_data[field] = value
_page = self._cache.get_by_natural_key(asset['natural_key'])
try:

View File

@@ -150,21 +150,19 @@ class Base(Page):
HTTPBasicAuth(client_id, client_secret)(req)
req.headers['Content-Type'] = 'application/x-www-form-urlencoded'
resp = self.connection.post(
f"{config.api_base_path}o/token/",
data={"grant_type": "password", "username": username, "password": password, "scope": scope},
headers=req.headers,
'/api/o/token/', data={"grant_type": "password", "username": username, "password": password, "scope": scope}, headers=req.headers
)
elif client_id:
req.headers['Content-Type'] = 'application/x-www-form-urlencoded'
resp = self.connection.post(
f"{config.api_base_path}o/token/",
'/api/o/token/',
data={"grant_type": "password", "username": username, "password": password, "client_id": client_id, "scope": scope},
headers=req.headers,
)
else:
HTTPBasicAuth(username, password)(req)
resp = self.connection.post(
'{0}v2/users/{1}/personal_tokens/'.format(config.api_base_path, username),
'/api/v2/users/{}/personal_tokens/'.format(username),
json={"description": description, "application": None, "scope": scope},
headers=req.headers,
)
@@ -209,7 +207,7 @@ class Base(Page):
jobs = []
for active_job in active_jobs:
job_type = active_job['type']
endpoint = '{}v2/{}s/{}/'.format(config.api_base_path, job_type, active_job['id'])
endpoint = '/api/v2/{}s/{}/'.format(job_type, active_job['id'])
job = self.walk(endpoint)
jobs.append(job)
job.cancel()

View File

@@ -1,6 +1,3 @@
from awxkit.config import config
class Resources(object):
_activity = r'activity_stream/\d+/'
_activity_stream = 'activity_stream/'
@@ -284,7 +281,7 @@ class Resources(object):
_workflow_job_workflow_nodes = r'workflow_jobs/\d+/workflow_nodes/'
_subscriptions = 'config/subscriptions/'
_workflow_jobs = 'workflow_jobs/'
api = str(config.api_base_path)
api = '/api/'
common = api + r'v\d+/'
v2 = api + 'v2/'

View File

@@ -15,12 +15,7 @@ def freeze(key):
def parse_description(desc):
options = {}
desc_lines = []
if 'POST' in desc:
desc_lines = desc[desc.index('POST') :].splitlines()
else:
desc_lines = desc.splitlines()
for line in desc_lines:
for line in desc[desc.index('POST') :].splitlines():
match = descRE.match(line)
if not match:
continue

View File

@@ -122,5 +122,5 @@ def as_user(v, username, password=None):
def uses_sessions(connection):
session_login = connection.get(f"{config.api_base_path}login/")
session_login = connection.get('/api/login/')
return session_login.status_code == 200

View File

@@ -4,7 +4,6 @@ import json
from .stdout import monitor, monitor_workflow
from .utils import CustomRegistryMeta, color_enabled
from awxkit import api
from awxkit.config import config
from awxkit.exceptions import NoContent
@@ -480,7 +479,7 @@ class RoleMixin(object):
options = ', '.join(RoleMixin.roles[flag])
raise ValueError("invalid choice: '{}' must be one of {}".format(role, options))
value = kwargs[flag]
target = '{}v2/{}/{}'.format(config.api_base_path, resource, value)
target = '/api/v2/{}/{}'.format(resource, value)
detail = self.page.__class__(target, self.page.connection).get()
object_roles = detail['summary_fields']['object_roles']
actual_role = object_roles[role + '_role']

View File

@@ -6,7 +6,6 @@ import sys
import time
from .utils import cprint, color_enabled, STATUS_COLORS
from awxkit.config import config
from awxkit.utils import to_str
@@ -18,7 +17,7 @@ def monitor_workflow(response, session, print_stdout=True, action_timeout=None,
}
def fetch(seen):
results = response.connection.get(f"{config.api_base_path}v2/unified_jobs", payload).json()['results']
results = response.connection.get('/api/v2/unified_jobs', payload).json()['results']
# erase lines we've previously printed
if print_stdout and sys.stdout.isatty():

View File

@@ -32,4 +32,3 @@ config.assume_untrusted = config.get('assume_untrusted', True)
config.client_connection_attempts = int(os.getenv('AWXKIT_CLIENT_CONNECTION_ATTEMPTS', 5))
config.prevent_teardown = to_bool(os.getenv('AWXKIT_PREVENT_TEARDOWN', False))
config.use_sessions = to_bool(os.getenv('AWXKIT_SESSIONS', False))
config.api_base_path = os.getenv('AWXKIT_API_BASE_PATH', '/api/')

View File

@@ -14,8 +14,10 @@ import yaml
from awxkit.words import words
from awxkit.exceptions import WaitUntilTimeout
log = logging.getLogger(__name__)
cloud_types = (
'aws',
'azure',

View File

@@ -90,7 +90,6 @@ setup(
install_requires=[
'PyYAML',
'requests',
'setuptools',
],
python_requires=">=3.8",
extras_require={'formatting': ['jq'], 'websockets': ['websocket-client==0.57.0'], 'crypto': ['cryptography']},

View File

@@ -16,11 +16,11 @@ pubdate = datetime.strptime(pubdateshort, '%Y-%m-%d').strftime('%B %d, %Y')
# The name for this set of Sphinx documents. If None, it defaults to
# "<project> v<release> documentation".
# html_title = None
#html_title = None
html_title = 'Ansible AWX community documentation'
# A shorter title for the navigation bar. Default is the same as html_title.
# html_short_title = None
#html_short_title = None
html_short_title = 'AWX community documentation'
htmlhelp_basename = 'AWX_docs'
@@ -54,8 +54,8 @@ release = 'AWX latest'
language = 'en'
locale_dirs = ['locale/'] # path is example but recommended.
gettext_compact = False # optional.
locale_dirs = ['locale/'] # path is example but recommended.
gettext_compact = False # optional.
rst_epilog = """
.. |atqi| replace:: *AWX Quick Installation Guide*
@@ -88,8 +88,4 @@ rst_epilog = """
.. |rhaap| replace:: Red Hat Ansible Automation Platform
.. |RHAT| replace:: Red Hat Ansible Automation Platform controller
""" % (
version,
pubdateshort,
pubdate,
)
""" % (version, pubdateshort, pubdate)

View File

@@ -51,213 +51,6 @@ Prerequisites
- To manage instances from the AWX user interface, you must have System Administrator or System Auditor permissions.
Common topologies
------------------
Instances make up the network of devices that communicate with one another. They are the building blocks of an automation mesh. These building blocks serve as nodes in a mesh topology. There are several kinds of instances:
+-----------+-----------------------------------------------------------------------------------------------------------------+
| Node Type | Description |
+===========+=================================================================================================================+
| Control | Nodes that run persistent Ansible Automation Platform services, and delegate jobs to hybrid and execution nodes |
+-----------+-----------------------------------------------------------------------------------------------------------------+
| Hybrid | Nodes that run persistent Ansible Automation Platform services and execute jobs |
| | (not applicable to operator-based installations) |
+-----------+-----------------------------------------------------------------------------------------------------------------+
| Hop | Used for relaying across the mesh only |
+-----------+-----------------------------------------------------------------------------------------------------------------+
| Execution | Nodes that run jobs delivered from control nodes (jobs submitted from the users Ansible automation) |
+-----------+-----------------------------------------------------------------------------------------------------------------+
Simple topology
~~~~~~~~~~~~~~~~
One of the ways to expand job capacity is to create a standalone execution node that can be added to run alongside the Kubernetes deployment of AWX. These machines will not be a part of the AWX Kubernetes cluster. The control nodes running in the cluster will connect and submit work to these machines via Receptor. The machines are registered in AWX as type "execution" instances, meaning they will only be used to run AWX jobs, not dispatch work or handle web requests as control nodes do.
Hop nodes can be added to sit between the control plane of AWX and standalone execution nodes. These machines will not be a part of the AWX Kubernetes cluster and they will be registered in AWX as node type "hop", meaning they will only handle inbound and outbound traffic for otherwise unreachable nodes in a different or more strict network.
Below is an example of an AWX task pod with two execution nodes. Traffic to execution node 2 flows through a hop node that is setup between it and the control plane.
.. image:: ../common/images/instances_awx_task_pods_hopnode.png
:alt: AWX task pod with a hop node between the control plane of AWX and standalone execution nodes.
An example of a simple topology may look like the following:
.. list-table::
:widths: 20 30 10 20 15
:header-rows: 1
* - Instance type
- Hostname
- Listener port
- Peers from control nodes
- Peers
* - Control plane
- awx-task-65d6d96987-mgn9j
- 27199
- True
- []
* - Hop node
- awx-hop-node
- 27199
- True
- []
* - Execution node
- awx-example.com
- n/a
- False
- ["hop node"]
Mesh topology
~~~~~~~~~~~~~~
Mesh ingress is a feature that allows remote nodes to connect inbound to the control plane. This is especially useful when creating remote nodes in restricted networking environments that disallow inbound traffic.
.. image:: ../common/images/instances_mesh_ingress_topology.png
:alt: Mesh ingress architecture showing the peering relationship between nodes.
An example of a topology that uses mesh ingress may look like the following:
.. list-table::
:widths: 20 30 10 20 15
:header-rows: 1
* - Instance type
- Hostname
- Listener port
- Peers from control nodes
- Peers
* - Control plane
- awx-task-xyz
- 27199
- True
- []
* - Hop node
- awx-hop-node
- 27199
- True
- []
* - Execution node
- awx-example.com
- n/a
- False
- ["hop node"]
In order to create a mesh ingress for AWX, see the `Mesh Ingress <https://ansible.readthedocs.io/projects/awx-operator/en/latest/user-guide/advanced-configuration/mesh-ingress.html>`_ chapter of the AWX Operator Documentation for information on setting up this type of topology. The last step is to create a remote execution node and add the execution node to an instance group in order for it to be used in your job execution. Whatever execution environment image used to run a playbook needs to be accessible for your remote execution node. Everything you are using in your playbook also needs to be accessible from this remote execution node.
.. image:: ../common/images/instances-job-template-using-remote-execution-ig.png
:alt: Job template using the instance group with the execution node to run jobs.
.. _ag_instances_add:
Add an instance
----------------
To create an instance in AWX:
1. Click **Instances** from the left side navigation menu of the AWX UI.
2. In the Instances list view, click the **Add** button and the Create new Instance window opens.
.. image:: ../common/images/instances_create_new.png
:alt: Create a new instance form.
An instance has several attributes that may be configured:
- Enter a fully qualified domain name (ping-able DNS) or IP address for your instance in the **Host Name** field (required). This field is equivalent to ``hostname`` in the API.
- Optionally enter a **Description** for the instance
- The **Instance State** field is auto-populated, indicating that it is being installed, and cannot be modified
- Optionally specify the **Listener Port** for the receptor to listen on for incoming connections. This is an open port on the remote machine used to establish inbound TCP connections. This field is equivalent to ``listener_port`` in the API.
- Select from the options in **Instance Type** field to specify the type you want to create. Only execution and hop nodes can be created as operator-based installations do not support hybrid nodes. This field is equivalent to ``node_type`` in the API.
- In the **Peers** field, select the instance hostnames you want your new instance to connect outbound to.
- In the **Options** fields:
- Check the **Enable Instance** box to make it available for jobs to run on an execution node.
- Check the **Managed by Policy** box to allow policy to dictate how the instance is assigned.
- Check the **Peers from control nodes** box to allow control nodes to peer to this instance automatically. Listener port needs to be set if this is enabled or the instance is a peer.
3. Once the attributes are configured, click **Save** to proceed.
Upon successful creation, the Details of the one of the created instances opens.
.. image:: ../common/images/instances_create_details.png
:alt: Details of the newly created instance.
.. note::
The proceeding steps 4-8 are intended to be ran from any computer that has SSH access to the newly created instance.
4. Click the download button next to the **Install Bundle** field to download the tarball that contain files to allow AWX to make proper TCP connections to the remote machine.
.. image:: ../common/images/instances_install_bundle.png
:alt: Instance details showing the Download button in the Install Bundle field of the Details tab.
5. Extract the downloaded ``tar.gz`` file from the location you downloaded it. The install bundle contains TLS certificates and keys, a certificate authority, and a proper Receptor configuration file. To facilitate that these files will be in the right location on the remote machine, the install bundle includes an ``install_receptor.yml`` playbook. The playbook requires the Receptor collection which can be obtained via:
::
ansible-galaxy collection install -r requirements.yml
6. Before running the ``ansible-playbook`` command, edit the following fields in the ``inventory.yml`` file:
- ``ansible_user`` with the username running the installation
- ``ansible_ssh_private_key_file`` to contain the filename of the private key used to connect to the instance
::
---
all:
hosts:
remote-execution:
ansible_host: <hostname>
ansible_user: <username> # user provided
ansible_ssh_private_key_file: ~/.ssh/id_rsa
The content of the ``inventory.yml`` file serves as a template and contains variables for roles that are applied during the installation and configuration of a receptor node in a mesh topology. You may modify some of the other fields, or replace the file in its entirety for advanced scenarios. Refer to `Role Variables <https://github.com/ansible/receptor-collection/blob/main/README.md>`_ for more information on each variable.
7. Save the file to continue.
8. Run the following command on the machine you want to update your mesh:
::
ansible-playbook -i inventory.yml install_receptor.yml
Wait a few minutes for the periodic AWX task to do a health check against the new instance. You may run a health check by selecting the node and clicking the **Run health check** button from its Details page at any time. Once the instances endpoint or page reports a "Ready" status for the instance, jobs are now ready to run on this machine!
9. To view other instances within the same topology or associate peers, click the **Peers** tab.
.. image:: ../common/images/instances_peers_tab.png
:alt: "Peers" tab showing two peers.
To associate peers with your node, click the **Associate** button to open a dialog box of instances eligible for peering.
.. image:: ../common/images/instances_associate_peer.png
:alt: Instances available to peer with the example hop node.
Execution nodes can peer with either hop nodes or other execution nodes. Hop nodes can only peer with execution nodes unless you check the **Peers from control nodes** check box from the **Options** field.
.. note::
If you associate or disassociate a peer, a notification will inform you to re-run the install bundle from the Peer Detail view (the :ref:`ag_topology_viewer` has the download link).
.. image:: ../common/images/instances_associate_peer_reinstallmsg.png
:alt: Notification to re-run the installation bundle due to change in the peering.
You can remove an instance by clicking **Remove** in the Instances page, or by setting the instance ``node_state = deprovisioning`` via the API. Upon deleting, a pop-up message will appear to notify that you may need to re-run the install bundle to make sure things that were removed are no longer connected.
10. To view a graphical representation of your updated topology, refer to the :ref:`ag_topology_viewer` section of this guide.
Manage instances
-----------------
@@ -320,10 +113,153 @@ The example health check shows the status updates with an error on node 'one':
:alt: Health check showing an error in one of the instances.
Add an instance
----------------
One of the ways to expand capacity is to create an instance. Standalone execution nodes can be added to run alongside the Kubernetes deployment of AWX. These machines will not be a part of the AWX Kubernetes cluster. The control nodes running in the cluster will connect and submit work to these machines via Receptor. The machines are registered in AWX as type "execution" instances, meaning they will only be used to run AWX jobs, not dispatch work or handle web requests as control nodes do.
Hop nodes can be added to sit between the control plane of AWX and standalone execution nodes. These machines will not be a part of the AWX Kubernetes cluster and they will be registered in AWX as node type "hop", meaning they will only handle inbound and outbound traffic for otherwise unreachable nodes in a different or more strict network.
Below is an example of an AWX task pod with two execution nodes. Traffic to execution node 2 flows through a hop node that is setup between it and the control plane.
.. image:: ../common/images/instances_awx_task_pods_hopnode.png
:alt: AWX task pod with a hop node between the control plane of AWX and standalone execution nodes.
To create an instance in AWV:
1. Click **Instances** from the left side navigation menu of the AWX UI.
2. In the Instances list view, click the **Add** button and the Create new Instance window opens.
.. image:: ../common/images/instances_create_new.png
:alt: Create a new instance form.
An instance has several attributes that may be configured:
- Enter a fully qualified domain name (ping-able DNS) or IP address for your instance in the **Host Name** field (required). This field is equivalent to ``hostname`` in the API.
- Optionally enter a **Description** for the instance
- The **Instance State** field is auto-populated, indicating that it is being installed, and cannot be modified
- Optionally specify the **Listener Port** for the receptor to listen on for incoming connections. This is an open port on the remote machine used to establish inbound TCP connections. This field is equivalent to ``listener_port`` in the API.
- Select from the options in **Instance Type** field to specify the type you want to create. Only execution and hop nodes can be created as operator-based installations do not support hybrid nodes. This field is equivalent to ``node_type`` in the API.
- In the **Peers** field, select the instance hostnames you want your new instance to connect outbound to.
- In the **Options** fields:
- Check the **Enable Instance** box to make it available for jobs to run on an execution node.
- Check the **Managed by Policy** box to allow policy to dictate how the instance is assigned.
- Check the **Peers from control nodes** box to allow control nodes to peer to this instance automatically. Listener port needs to be set if this is enabled or the instance is a peer.
In the example diagram above, the configurations are as follows:
+------------------+---------------+--------------------------+--------------+
| instance name | listener_port | peers_from_control_nodes | peers |
+==================+===============+==========================+==============+
| execution node 1 | 27199 | true | [] |
+------------------+---------------+--------------------------+--------------+
| hop node | 27199 | true | [] |
+------------------+---------------+--------------------------+--------------+
| execution node 2 | null | false | ["hop node"] |
+------------------+---------------+--------------------------+--------------+
3. Once the attributes are configured, click **Save** to proceed.
Upon successful creation, the Details of the one of the created instances opens.
.. image:: ../common/images/instances_create_details.png
:alt: Details of the newly created instance.
.. note::
The proceeding steps 4-8 are intended to be ran from any computer that has SSH access to the newly created instance.
4. Click the download button next to the **Install Bundle** field to download the tarball that contain files to allow AWX to make proper TCP connections to the remote machine.
.. image:: ../common/images/instances_install_bundle.png
:alt: Instance details showing the Download button in the Install Bundle field of the Details tab.
5. Extract the downloaded ``tar.gz`` file from the location you downloaded it. The install bundle contains TLS certificates and keys, a certificate authority, and a proper Receptor configuration file. To facilitate that these files will be in the right location on the remote machine, the install bundle includes an ``install_receptor.yml`` playbook. The playbook requires the Receptor collection which can be obtained via:
::
ansible-galaxy collection install -r requirements.yml
6. Before running the ``ansible-playbook`` command, edit the following fields in the ``inventory.yml`` file:
- ``ansible_user`` with the username running the installation
- ``ansible_ssh_private_key_file`` to contain the filename of the private key used to connect to the instance
::
---
all:
hosts:
remote-execution:
ansible_host: 18.206.206.34
ansible_user: <username> # user provided
ansible_ssh_private_key_file: ~/.ssh/id_rsa
The content of the ``inventory.yml`` file serves as a template and contains variables for roles that are applied during the installation and configuration of a receptor node in a mesh topology. You may modify some of the other fields, or replace the file in its entirety for advanced scenarios. Refer to `Role Variables <https://github.com/ansible/receptor-collection/blob/main/README.md>`_ for more information on each variable.
7. Save the file to continue.
8. Run the following command on the machine you want to update your mesh:
::
ansible-playbook -i inventory.yml install_receptor.yml
Wait a few minutes for the periodic AWX task to do a health check against the new instance. You may run a health check by selecting the node and clicking the **Run health check** button from its Details page at any time. Once the instances endpoint or page reports a "Ready" status for the instance, jobs are now ready to run on this machine!
9. To view other instances within the same topology or associate peers, click the **Peers** tab.
.. image:: ../common/images/instances_peers_tab.png
:alt: "Peers" tab showing two peers.
To associate peers with your node, click the **Associate** button to open a dialog box of instances eligible for peering.
.. image:: ../common/images/instances_associate_peer.png
:alt: Instances available to peer with the example hop node.
Execution nodes can peer with either hop nodes or other execution nodes. Hop nodes can only peer with execution nodes unless you check the **Peers from control nodes** check box from the **Options** field.
.. note::
If you associate or disassociate a peer, a notification will inform you to re-run the install bundle from the Peer Detail view (the :ref:`ag_topology_viewer` has the download link).
.. image:: ../common/images/instances_associate_peer_reinstallmsg.png
:alt: Notification to re-run the installation bundle due to change in the peering.
You can remove an instance by clicking **Remove** in the Instances page, or by setting the instance ``node_state = deprovisioning`` via the API. Upon deleting, a pop-up message will appear to notify that you may need to re-run the install bundle to make sure things that were removed are no longer connected.
10. To view a graphical representation of your updated topology, refer to the :ref:`ag_topology_viewer` section of this guide.
Using a custom Receptor CA
---------------------------
Refer to the AWX Operator Documentation, `Custom Receptor CA <https://ansible.readthedocs.io/projects/awx-operator/en/latest/user-guide/advanced-configuration/custom-receptor-certs.html>`_ for detail.
The control nodes on the K8S cluster will communicate with execution nodes via mutual TLS TCP connections, running via Receptor. Execution nodes will verify incoming connections by ensuring the x509 certificate was issued by a trusted Certificate Authority (CA).
You may choose to provide your own CA for this validation. If no CA is provided, AWX operator will automatically generate one using OpenSSL.
Given custom ``ca.crt`` and ``ca.key`` stored locally, run the following:
::
kubectl create secret tls awx-demo-receptor-ca \
--cert=/path/to/ca.crt --key=/path/to/ca.key
The secret should be named ``{AWX Custom Resource name}-receptor-ca``. In the above, the AWX Custom Resource name is "awx-demo". Replace "awx-demo" with your AWX Custom Resource name.
If this secret is created after AWX is deployed, run the following to restart the deployment:
::
kubectl rollout restart deployment awx-demo
.. note::
Changing the receptor CA will sever connections to any existing execution nodes. These nodes will enter an *Unavailable* state, and jobs will not be able to run on them. You will need to download and re-run the install bundle for each execution node. This will replace the TLS certificate files with those signed by the new CA. The execution nodes will then appear in a *Ready* state after a few minutes.
Using a private image for the default EE

Binary file not shown.

Before

Width:  |  Height:  |  Size: 128 KiB

After

Width:  |  Height:  |  Size: 78 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 130 KiB

After

Width:  |  Height:  |  Size: 207 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 49 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 3.6 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 102 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 91 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 134 KiB

After

Width:  |  Height:  |  Size: 43 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 52 KiB

After

Width:  |  Height:  |  Size: 66 KiB

Some files were not shown because too many files have changed in this diff Show More