openstack
/
tacker
Code Issues Proposed changes

Move usage from deprecated Ceilometer API to Gnocchi API 

This removes usage of deprecated Ceilometer API

1. Changing trigger type from "OS::Ceilometer::Alarm"
to "OS::Aodh::GnocchiAggregationByResourcesAlarm"
2. Add "resource_type" fixed with value "instance"
3. Change some paramters (meter_name-> metric,
statistic ->aggregation_method, period-> granularity)
4. Change value from "average" to "mean" in method to
compare to the threshold

Change-Id: I486c14cbc9d05a0e826bbef1ad181bdcb2d8c951
Closes-Bug: #1735484
This commit is contained in:
hoangphuocbk 2017-12-01 02:37:38 +09:00 committed by Cong Phuoc Hoang
parent 7a0efa8007
commit 0da9469017
25 changed files with 285 additions and 249 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

169
doc/source/user/alarm_monitoring_usage_guide.rst
View File

@ -28,23 +28,25 @@ described firstly like other TOSCA templates in Tacker.
.. code-block:: yaml
policies:
- vdu1_cpu_usage_monitoring_policy:
type: tosca.policies.tacker.Alarming
triggers:
resize_compute:
policies:
- vdu1_cpu_usage_monitoring_policy:
type: tosca.policies.tacker.Alarming
triggers:
vdu_hcpu_usage_respawning:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
metrics: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 65
granularity: 600
evaluations: 1
method: avg
aggregation_method: mean
resource_type: instance
comparison_operator: gt
actions: [respawn]
metadata: VDU1
action: [respawn]
Alarm framework already supported the some default backend actions like
**scaling, respawn, log, and log_and_kill**.
@ -77,7 +79,7 @@ in Tacker:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -89,30 +91,6 @@ in Tacker:
node: VL1
- virtualBinding:
node: VDU1
VDU2:
type: tosca.nodes.nfv.VDU.Tacker
capabilities:
nfv_compute:
properties:
disk_size: 1 GB
mem_size: 512 MB
num_cpus: 2
properties:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
CP2:
type: tosca.nodes.nfv.CP.Tacker
properties:
management: true
anti_spoofing_protection: false
requirements:
- virtualLink:
node: VL1
- virtualBinding:
node: VDU2
VL1:
type: tosca.nodes.nfv.VL
@ -123,13 +101,13 @@ in Tacker:
policies:
- SP1:
type: tosca.policies.tacker.Scaling
targets: [VDU1]
properties:
increment: 1
cooldown: 120
min_instances: 1
max_instances: 3
default_instances: 2
targets: [VDU1,VDU2]
default_instances: 1
- vdu_cpu_usage_monitoring_policy:
type: tosca.policies.tacker.Alarming
@ -138,32 +116,33 @@ in Tacker:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
metrics: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
threshold: 80
constraint: utilization greater_than 80%
granularity: 300
evaluations: 1
method: avg
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: SG1
actions: [SP1]
action: [SP1]
vdu_lcpu_usage_scaling_in:
targets: [VDU1, VDU2]
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
metrics: cpu_util
metric: cpu_util
condition:
threshold: 10
constraint: utilization less_than 10%
period: 600
granularity: 300
evaluations: 1
method: avg
aggregation_method: mean
resource_type: instance
comparison_operator: lt
metadata: SG1
actions: [SP1]
action: [SP1]
**NOTE:**
@ -188,6 +167,22 @@ How to monitor VNFs via alarm triggers
How to setup alarm configuration
================================
Tacker provides templates that implemented Ceilometer as alarm for monitoring
VNFs, which are located in **tacker/samples/tosca-templates/vnfd**.
1. tosca-vnfd-alarm-multi-actions.yaml
2. tosca-vnfd-alarm-respawn.yaml
3. tosca-vnfd-alarm-scale.yaml
The following commands shows creating VNF with alarms for scaling in and out.
.. code-block:: console
$ cd ~/tacker/samples/tosca-templates/vnfd
$ openstack vnf create --vnfd-template tosca-vnfd-alarm-scale.yaml VNF1
Firstly, vnfd and vnf need to be created successfully using pre-defined TOSCA
template for alarm monitoring. Then, in order to know whether alarm
configuration defined in Tacker is successfully passed to Ceilometer,
@ -195,51 +190,53 @@ Tacker users could use CLI:
.. code-block:: console
$ aodh alarm list
$ openstack alarm list
+--------------------------------------+--------------------------------------------+-----------------------------------------------------------------------------------+-------------------+----------+---------+
| alarm_id | type | name | state | severity | enabled |
+--------------------------------------+--------------------------------------------+-----------------------------------------------------------------------------------+-------------------+----------+---------+
| f418ebf8-f8a6-4991-8f0d-938e38434411 | gnocchi_aggregation_by_resources_threshold | VNF1_7582cdf4-58ed-4df8-8fa2-c15938adf70b-vdu_hcpu_usage_scaling_out-4imzw3c7cicb | insufficient data | low | True |
| 70d86622-940a-4bc3-87c2-d5dfbb01bbea | gnocchi_aggregation_by_resources_threshold | VNF1_7582cdf4-58ed-4df8-8fa2-c15938adf70b-vdu_lcpu_usage_scaling_in-dwvdvbegiqdk | insufficient data | low | True |
+--------------------------------------+--------------------------------------------+-----------------------------------------------------------------------------------+-------------------+----------+---------+
+--------------------------------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------+-------------------+----------+---------+
| alarm_id | type | name | state | severity | enabled |
+--------------------------------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------+-------------------+----------+---------+
| 6f2336b9-e0a2-4e33-88be-bc036192b42b | threshold | tacker.vnfm.infra_drivers.openstack.openstack_OpenStack-a0f60b00-ad3d-4769-92ef-e8d9518da2c8-vdu_lcpu_scaling_in-smgctfnc3ql5 | insufficient data | low | True |
| e049f0d3-09a8-46c0-9b88-e61f1f524aab | threshold | tacker.vnfm.infra_drivers.openstack.openstack_OpenStack-a0f60b00-ad3d-4769-92ef-e8d9518da2c8-vdu_hcpu_usage_scaling_out-lubylov5g6xb | insufficient data | low | True |
+--------------------------------------+-----------+--------------------------------------------------------------------------------------------------------------------------------------+-------------------+----------+---------+
.. code-block:: console
$aodh alarm show 6f2336b9-e0a2-4e33-88be-bc036192b42b
+---------------------------+-------------------------------------------------------------------------------------------------------------------------------+
| Field | Value |
+---------------------------+-------------------------------------------------------------------------------------------------------------------------------+
| alarm_actions | [u'http://pinedcn:9890/v1.0/vnfs/a0f60b00-ad3d-4769-92ef-e8d9518da2c8/vdu_lcpu_scaling_in/SP1-in/yl7kh5qd'] |
| alarm_id | 6f2336b9-e0a2-4e33-88be-bc036192b42b |
| comparison_operator | lt |
| description | utilization less_than 10% |
| enabled | True |
| evaluation_periods | 1 |
| exclude_outliers | False |
| insufficient_data_actions | None |
| meter_name | cpu_util |
| name | tacker.vnfm.infra_drivers.openstack.openstack_OpenStack-a0f60b00-ad3d-4769-92ef-e8d9518da2c8-vdu_lcpu_scaling_in-smgctfnc3ql5 |
| ok_actions | None |
| period | 600 |
| project_id | 3db801789c9e4b61b14ce448c9e7fb6d |
| query | metadata.user_metadata.vnf_id = a0f60b00-ad3d-4769-92ef-e8d9518da2c8 |
| repeat_actions | True |
| severity | low |
| state | insufficient data |
| state_timestamp | 2016-11-16T18:39:30.134954 |
| statistic | avg |
| threshold | 10.0 |
| time_constraints | [] |
| timestamp | 2016-11-16T18:39:30.134954 |
| type | threshold |
| user_id | a783e8a94768484fb9a43af03c6426cb |
+---------------------------+-------------------------------------------------------------------------------------------------------------------------------+
$ openstack alarm show 70d86622-940a-4bc3-87c2-d5dfbb01bbea
+---------------------------+------------------------------------------------------------------------------------------------------------------+
| Field | Value |
+---------------------------+------------------------------------------------------------------------------------------------------------------+
| aggregation_method | mean |
| alarm_actions | [u'http://ubuntu:9890/v1.0/vnfs/7582cdf4-58ed-4df8-8fa2-c15938adf70b/vdu_lcpu_usage_scaling_in/SP1-in/v2fq7rd7'] |
| alarm_id | 70d86622-940a-4bc3-87c2-d5dfbb01bbea |
| comparison_operator | lt |
| description | utilization less_than 10% |
| enabled | True |
| evaluation_periods | 1 |
| granularity | 60 |
| insufficient_data_actions | [] |
| metric | cpu_util |
| name | VNF1_7582cdf4-58ed-4df8-8fa2-c15938adf70b-vdu_lcpu_usage_scaling_in-dwvdvbegiqdk |
| ok_actions | [] |
| project_id | b5e054a3861b4da2b084aca9530096be |
| query | {"=": {"server_group": "SG1-64beb5e4-c0"}} |
| repeat_actions | True |
| resource_type | instance |
| severity | low |
| state | insufficient data |
| state_reason | Not evaluated yet |
| state_timestamp | 2018-07-20T06:00:33.142762 |
| threshold | 10.0 |
| time_constraints | [] |
| timestamp | 2018-07-20T06:00:33.142762 |
| type | gnocchi_aggregation_by_resources_threshold |
| user_id | 61fb5c6193e549f3baee26bd508c0b29 |
+---------------------------+------------------------------------------------------------------------------------------------------------------+
How to trigger alarms:
======================
As shown in the above Ceilometer command, alarm state is shown as
"insufficient data". Alarm is triggered by Ceilometer once alarm
state changes to "alarm".
@ -252,9 +249,9 @@ in **/etc/ceilometer/pipeline.yaml** file and then restart Ceilometer service.
Another way could be used to check if backend action is handled well in Tacker:
.. code-block::ini
.. code-block:: console
curl -H "Content-Type: application/json" -X POST -d '{"alarm_id": "35a80852-e24f-46ed-bd34-e2f831d00172", "current": "alarm"}' http://pinedcn:9890/v1.0/vnfs/a0f60b00-ad3d-4769-92ef-e8d9518da2c8/vdu_lcpu_scaling_in/SP1-in/yl7kh5qd
curl -H "Content-Type: application/json" -X POST -d '{"alarm_id": "35a80852-e24f-46ed-bd34-e2f831d00172", "current": "alarm"}' http://ubuntu:9890/v1.0/vnfs/7582cdf4-58ed-4df8-8fa2-c15938adf70b/vdu_lcpu_usage_scaling_in/SP1-in/v2fq7rd7
Then, users can check Horizon to know if vnf is respawned. Please note
that the url used in the above command could be captured from

11
releasenotes/notes/move-usage-from-ceilometer-to-gnocchi-api-a08d47042428ec0a.yaml Normal file
View File

@ -0,0 +1,11 @@
---
fixes:
- |
Removes usage of deprecated Ceilometer API:
1. Change trigger type from "OS::Ceilometer::Alarm" to
"OS::Aodh::GnocchiAggregationByResourcesAlarm"
2. Add "resource_type" fixed with value "instance"
3. Change some paramters (meter_name-> metric, statistic ->
aggregation_method, period-> granularity)
4. Change value from "average" to "mean" in method to compare
to the threshold

9
samples/tosca-templates/vnfd/tosca-vnfd-alarm-multi-actions.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -45,13 +45,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
comparison_operator: gt
resource_type: instance
metadata: VDU1
action: [respawn, log]

9
samples/tosca-templates/vnfd/tosca-vnfd-alarm-respawn.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -45,13 +45,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 300
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
action: [respawn]

48
samples/tosca-templates/vnfd/tosca-vnfd-alarm-scale.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -30,30 +30,6 @@ topology_template:
node: VL1
- virtualBinding:
node: VDU1
VDU2:
type: tosca.nodes.nfv.VDU.Tacker
capabilities:
nfv_compute:
properties:
disk_size: 1 GB
mem_size: 512 MB
num_cpus: 2
properties:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
CP2:
type: tosca.nodes.nfv.CP.Tacker
properties:
management: true
anti_spoofing_protection: false
requirements:
- virtualLink:
node: VL1
- virtualBinding:
node: VDU2
VL1:
type: tosca.nodes.nfv.VL
@ -64,13 +40,13 @@ topology_template:
policies:
- SP1:
type: tosca.policies.tacker.Scaling
targets: [VDU1,VDU2]
targets: [VDU1]
properties:
increment: 1
cooldown: 120
min_instances: 1
max_instances: 3
default_instances: 2
default_instances: 1
- vdu_cpu_usage_monitoring_policy:
type: tosca.policies.tacker.Alarming
@ -79,13 +55,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
threshold: 80
constraint: utilization greater_than 80%
granularity: 60
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: SG1
action: [SP1]
@ -94,13 +71,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 10
constraint: utilization less_than 10%
period: 600
granularity: 60
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: lt
metadata: SG1
action: [SP1]

5
tacker/extensions/vnfm.py
View File

@ -149,6 +149,11 @@ class MetadataNotMatched(exceptions.InvalidInput):
message = _("Metadata for alarm policy is not matched")
class InvalidResourceType(exceptions.InvalidInput):
message = _("Resource type %(resource_type)s for alarm policy "
"is not supported")
class InvalidSubstitutionMapping(exceptions.InvalidInput):
message = _("Input for substitution mapping requirements are not"
" valid for %(requirement)s. They must be in the form"

9
tacker/tests/etc/samples/sample-tosca-alarm-respawn.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -45,13 +45,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
action: [respawn]

16
tacker/tests/etc/samples/sample-tosca-alarm-scale.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -55,13 +55,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: SG1
action: [SP1]
@ -70,13 +71,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 10
constraint: utilization less_than 10%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: lt
metadata: SG1
action: [SP1]

3
tacker/tests/functional/vnfm/test_tosca_vnf_alarm.py
View File

@ -13,7 +13,6 @@
# under the License.
import json
import time
import unittest
from tacker.plugins.common import constants as evt_constants
from tacker.tests import constants
@ -139,13 +138,11 @@ class VnfTestAlarmMonitor(base.BaseTackerTest):
self.addCleanup(self.wait_until_vnf_delete, vnf_id,
constants.VNF_CIRROS_DELETE_TIMEOUT)
@unittest.skip("Skip to wait for Heat-translator to support gnocchi alarm")
def test_vnf_alarm_respawn(self):
self._test_vnf_tosca_alarm(
'sample-tosca-alarm-respawn.yaml',
'alarm and respawn-vnf')
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_vnf_alarm_scale(self):
self._test_vnf_tosca_alarm(
'sample-tosca-alarm-scale.yaml',

6
tacker/tests/unit/test_tosca_templates_under_samples.py
View File

@ -14,13 +14,13 @@
import os
import testtools
import unittest
from tacker.tosca import utils
from toscaparser import tosca_template
from toscaparser.utils import yamlparser
from translator.hot import tosca_translator
from tacker.tosca import utils
class TestSamples(testtools.TestCase):
"""Sample tosca validation.
@ -82,11 +82,9 @@ class TestSamples(testtools.TestCase):
def test_scale_sample(self, tosca_file=['tosca-vnfd-scale.yaml']):
self._test_samples(tosca_file)
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_alarm_sample(self, tosca_file=['tosca-vnfd-alarm-scale.yaml']):
self._test_samples(tosca_file)
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_list_samples(self,
files=['tosca-vnfd-scale.yaml',
'tosca-vnfd-alarm-scale.yaml']):

2
tacker/tests/unit/vnfm/infra_drivers/openstack/data/hot_alarm_scale_custom.yaml
View File

@ -14,7 +14,7 @@ resources:
- port: { get_resource: CP1 }
image: cirros-0.4.0-x86_64-disk
flavor: m1.tiny
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1-2e3261d9-14}
VL1:
type: OS::Neutron::Net
CP1:

15
tacker/tests/unit/vnfm/infra_drivers/openstack/data/hot_tosca_alarm_metadata.yaml
View File

@ -29,13 +29,18 @@ resources:
ram: 512
vcpus: 2
vdu_hcpu_usage_respawning:
type: OS::Aodh::Alarm
type: OS::Aodh::GnocchiAggregationByResourcesAlarm
properties:
description: utilization greater_than 50%
meter_name: cpu_util
metric: cpu_util
threshold: 50
period: 60
statistic: avg
granularity: 60
aggregation_method: mean
resource_type: instance
evaluation_periods: 1
comparison_operator: gt
'matching_metadata': {'metadata.user_metadata.vnf': 'VDU1'}
query:
str_replace:
template: '{"=": {"server_group": "scaling_group_id"}}'
params:
scaling_group_id: VDU1-2e3261d9-1

17
tacker/tests/unit/vnfm/infra_drivers/openstack/data/hot_tosca_alarm_respawn.yaml
View File

@ -18,7 +18,7 @@ resources:
networks:
- port: {get_resource: CP1}
user_data_format: SOFTWARE_CONFIG
metadata: {'metering.vnf': 'VDU1'}
metadata: {'metering.server_group': VDU1-2e3261d9-1}
type: OS::Nova::Server
CP1:
properties: {network: net_mgmt, port_security_enabled: false}
@ -30,13 +30,18 @@ resources:
ram: 512
vcpus: 2
vdu_hcpu_usage_respawning:
type: OS::Aodh::Alarm
type: OS::Aodh::GnocchiAggregationByResourcesAlarm
properties:
description: utilization greater_than 50%
meter_name: cpu_util
metric: cpu_util
threshold: 50
period: 60
statistic: avg
granularity: 60
aggregation_method: mean
resource_type: instance
evaluation_periods: 1
comparison_operator: gt
'matching_metadata': {'metadata.user_metadata.vnf': 'VDU1'}
query:
str_replace:
template: '{"=": {"server_group": "scaling_group_id"}}'
params:
scaling_group_id: VDU1-2e3261d9-1

30
tacker/tests/unit/vnfm/infra_drivers/openstack/data/hot_tosca_alarm_scale.yaml
View File

@ -30,24 +30,34 @@ resources:
type: OS::Heat::ScalingPolicy
vdu_hcpu_usage_scaling_out:
type: OS::Aodh::Alarm
type: OS::Aodh::GnocchiAggregationByResourcesAlarm
properties:
description: utilization greater_than 50%
meter_name: cpu_util
statistic: avg
period: 600
metric: cpu_util
aggregation_method: mean
granularity: 600
evaluation_periods: 1
threshold: 50
matching_metadata: {'metadata.user_metadata.vnf': SG1}
resource_type: instance
query:
str_replace:
template: '{"=": {"server_group": "scaling_group_id"}}'
params:
scaling_group_id: SG1-2e3261d9-14
comparison_operator: gt
vdu_lcpu_usage_scaling_in:
type: OS::Aodh::Alarm
type: OS::Aodh::GnocchiAggregationByResourcesAlarm
properties:
description: utilization less_than 10%
meter_name: cpu_util
statistic: avg
period: 600
metric: cpu_util
aggregation_method: mean
granularity: 600
evaluation_periods: 1
threshold: 10
matching_metadata: {'metadata.user_metadata.vnf': SG1}
resource_type: instance
query:
str_replace:
template: '{"=": {"server_group": "scaling_group_id"}}'
params:
scaling_group_id: SG1-2e3261d9-14
comparison_operator: lt

11
tacker/tests/unit/vnfm/infra_drivers/openstack/data/test_tosca_vnfd_alarm_multi_actions.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1-2e3261d9-1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -45,13 +45,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
metadata: VDU1-2e3261d9-1
actions: [respawn, log]

11
tacker/tests/unit/vnfm/infra_drivers/openstack/data/test_tosca_vnfd_alarm_respawn.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1-2e3261d9-1}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -45,13 +45,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
metadata: VDU1-2e3261d9-1
action: [respawn]

11
tacker/tests/unit/vnfm/infra_drivers/openstack/data/test_tosca_vnfd_alarm_scale.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1-2e3261d9-14}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -55,13 +55,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: SG1
metadata: SG1-2e3261d9-14
action: [SP1]

10
tacker/tests/unit/vnfm/infra_drivers/openstack/data/tosca_alarm_metadata.yaml
View File

@ -22,7 +22,6 @@ topology_template:
mgmt_driver: noop
availability_zone: nova
CP1:
type: tosca.nodes.nfv.CP.Tacker
properties:
@ -48,13 +47,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: Ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 60
granularity: 60
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
metadata: VDU1-2e3261d9-1
action: ''

11
tacker/tests/unit/vnfm/infra_drivers/openstack/data/tosca_alarm_respawn.yaml
View File

@ -18,7 +18,7 @@ topology_template:
image: cirros-0.4.0-x86_64-disk
mgmt_driver: noop
availability_zone: nova
metadata: {metering.vnf: VDU1}
metadata: {metering.server_group: VDU1-2e3261d9-1}
CP1:
@ -46,13 +46,14 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: Ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 60
granularity: 60
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: VDU1
metadata: VDU1-2e3261d9-1
action: ''

20
tacker/tests/unit/vnfm/infra_drivers/openstack/data/tosca_alarm_scale.yaml
View File

@ -14,7 +14,7 @@ topology_template:
mgmt_driver: noop
availability_zone: nova
flavor: m1.tiny
metadata: {metering.vnf: SG1}
metadata: {metering.server_group: SG1-2e3261d9-14}
CP1:
type: tosca.nodes.nfv.CP.Tacker
@ -51,28 +51,30 @@ topology_template:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 50
constraint: utilization greater_than 50%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: gt
metadata: SG1
metadata: SG1-2e3261d9-14
action: [SP1]
vdu_lcpu_usage_scaling_in:
event_type:
type: tosca.events.resource.utilization
implementation: ceilometer
meter_name: cpu_util
metric: cpu_util
condition:
threshold: 10
constraint: utilization less_than 10%
period: 600
granularity: 600
evaluations: 1
method: average
aggregation_method: mean
resource_type: instance
comparison_operator: lt
metadata: SG1
metadata: SG1-2e3261d9-14
action: [SP1]

4
tacker/tests/unit/vnfm/infra_drivers/openstack/test_openstack.py
View File

@ -17,7 +17,6 @@ import codecs
import json
import mock
import os
import unittest
import yaml
from tacker import context
@ -447,7 +446,6 @@ class TestOpenStack(base.TestCase):
'hot_tosca_mac_ip.yaml'
)
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_create_tosca_alarm_respawn(self):
self._test_assert_equal_for_tosca_templates(
'tosca_alarm_respawn.yaml',
@ -455,7 +453,6 @@ class TestOpenStack(base.TestCase):
is_monitor=False
)
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_create_tosca_alarm_scale(self):
self._test_assert_equal_for_tosca_templates(
'tosca_alarm_scale.yaml',
@ -464,7 +461,6 @@ class TestOpenStack(base.TestCase):
is_monitor=False
)
@unittest.skip("Skip and wait for releasing Heat Translator")
def test_create_tosca_with_alarm_monitoring_not_matched(self):
self.assertRaises(vnfm.MetadataNotMatched,
self._test_assert_equal_for_tosca_templates,

81
tacker/tosca/utils.py
View File

@ -183,50 +183,67 @@ def get_vdu_applicationmonitoring(template):
@log.log
def get_vdu_metadata(template):
def get_vdu_metadata(template, unique_id=None):
metadata = dict()
metadata.setdefault('vdus', {})
for nt in template.nodetemplates:
if nt.type_definition.is_derived_from(TACKERVDU):
metadata_dict = nt.get_property_value('metadata') or None
if metadata_dict:
metadata_dict['metering.server_group'] = \
(metadata_dict['metering.server_group'] + '-'
+ unique_id)[:15]
metadata['vdus'][nt.name] = {}
metadata['vdus'][nt.name].update(metadata_dict)
return metadata
@log.log
def pre_process_alarm_resources(vnf, template, vdu_metadata):
def pre_process_alarm_resources(vnf, template, vdu_metadata, unique_id=None):
alarm_resources = dict()
matching_metadata = dict()
query_metadata = dict()
alarm_actions = dict()
for policy in template.policies:
if (policy.type_definition.is_derived_from(MONITORING)):
matching_metadata =\
_process_matching_metadata(vdu_metadata, policy)
if policy.type_definition.is_derived_from(MONITORING):
query_metadata = _process_query_metadata(
vdu_metadata, policy, unique_id)
alarm_actions = _process_alarm_actions(vnf, policy)
alarm_resources['matching_metadata'] = matching_metadata
alarm_resources['query_metadata'] = query_metadata
alarm_resources['alarm_actions'] = alarm_actions
return alarm_resources
def _process_matching_metadata(metadata, policy):
matching_mtdata = dict()
def _process_query_metadata(metadata, policy, unique_id):
query_mtdata = dict()
triggers = policy.entity_tpl['triggers']
for trigger_name, trigger_dict in triggers.items():
if not (trigger_dict.get('metadata') and metadata):
raise vnfm.MetadataNotMatched()
is_matched = False
for vdu_name, metadata_dict in metadata['vdus'].items():
if trigger_dict['metadata'] ==\
metadata_dict['metering.vnf']:
is_matched = True
if not is_matched:
raise vnfm.MetadataNotMatched()
matching_mtdata[trigger_name] = dict()
matching_mtdata[trigger_name]['metadata.user_metadata.vnf'] =\
trigger_dict['metadata']
return matching_mtdata
resource_type = trigger_dict.get('condition').get('resource_type')
# TODO(phuoc): currently, Tacker only supports resource_type with
# instance value. Other types such as instance_network_interface,
# instance_disk can be supported in the future.
if resource_type == 'instance':
if not (trigger_dict.get('metadata') and metadata):
raise vnfm.MetadataNotMatched()
is_matched = False
for vdu_name, metadata_dict in metadata['vdus'].items():
trigger_dict['metadata'] = \
(trigger_dict['metadata'] + '-' + unique_id)[:15]
if trigger_dict['metadata'] == \
metadata_dict['metering.server_group']:
is_matched = True
if not is_matched:
raise vnfm.MetadataNotMatched()
query_template = dict()
query_template['str_replace'] = dict()
query_template['str_replace']['template'] = \
'{"=": {"server_group": "scaling_group_id"}}'
scaling_group_param = \
{'scaling_group_id': trigger_dict['metadata']}
query_template['str_replace']['params'] = scaling_group_param
else:
raise vnfm.InvalidResourceType(resource_type=resource_type)
query_mtdata[trigger_name] = query_template
return query_mtdata
def _process_alarm_actions(vnf, policy):
@ -387,8 +404,8 @@ def represent_odict(dump, tag, mapping, flow_style=None):
@log.log
def post_process_heat_template(heat_tpl, mgmt_ports, metadata,
alarm_resources, res_tpl,
vol_res={}, unsupported_res_prop=None):
alarm_resources, res_tpl, vol_res={},
unsupported_res_prop=None, unique_id=None):
#
# TODO(bobh) - remove when heat-translator can support literal strings.
#
@ -412,19 +429,21 @@ def post_process_heat_template(heat_tpl, mgmt_ports, metadata,
LOG.debug('Added output for %s', outputname)
if metadata:
for vdu_name, metadata_dict in metadata['vdus'].items():
metadata_dict['metering.server_group'] = \
(metadata_dict['metering.server_group'] + '-' + unique_id)[:15]
if heat_dict['resources'].get(vdu_name):
heat_dict['resources'][vdu_name]['properties']['metadata'] =\
metadata_dict
matching_metadata = alarm_resources.get('matching_metadata')
query_metadata = alarm_resources.get('query_metadata')
alarm_actions = alarm_resources.get('alarm_actions')
if matching_metadata:
for trigger_name, matching_metadata_dict in matching_metadata.items():
if query_metadata:
for trigger_name, matching_metadata_dict in query_metadata.items():
if heat_dict['resources'].get(trigger_name):
matching_mtdata = dict()
matching_mtdata['matching_metadata'] =\
matching_metadata[trigger_name]
query_mtdata = dict()
query_mtdata['query'] = \
query_metadata[trigger_name]
heat_dict['resources'][trigger_name]['properties'].\
update(matching_mtdata)
update(query_mtdata)
if alarm_actions:
for trigger_name, alarm_actions_dict in alarm_actions.items():
if heat_dict['resources'].get(trigger_name):

6
tacker/vnfm/infra_drivers/openstack/openstack.py
View File

@ -335,6 +335,12 @@ class OpenStack(abstract_driver.DeviceAbstractDriver,
if events[0].id != last_event_id:
if events[0].resource_status == 'SIGNAL_COMPLETE':
break
else:
# When the number of instance reaches min or max, the below
# comparision will let VNF status turn into ACTIVE state.
if events[0].resource_status == 'CREATE_COMPLETE' or \
events[0].resource_status == 'SIGNAL_COMPLETE':
break
except Exception as e:
error_reason = _("VNF scaling failed for stack %(stack)s with "
"error %(error)s") % {

12
tacker/vnfm/infra_drivers/openstack/translate_template.py
View File

@ -13,6 +13,7 @@
from oslo_config import cfg
from oslo_log import log as logging
from oslo_serialization import jsonutils
from oslo_utils import uuidutils
from toscaparser import tosca_template
from toscaparser.utils import yamlparser
from translator.hot import tosca_translator
@ -275,9 +276,11 @@ class TOSCAToHOT(object):
LOG.debug("tosca-parser error: %s", str(e))
raise vnfm.ToscaParserFailed(error_msg_details=str(e))
metadata = toscautils.get_vdu_metadata(tosca)
alarm_resources =\
toscautils.pre_process_alarm_resources(self.vnf, tosca, metadata)
unique_id = uuidutils.generate_uuid()
metadata = toscautils.get_vdu_metadata(tosca, unique_id=unique_id)
alarm_resources = toscautils.pre_process_alarm_resources(
self.vnf, tosca, metadata, unique_id=unique_id)
monitoring_dict = toscautils.get_vdu_monitoring(tosca)
mgmt_ports = toscautils.get_mgmt_ports(tosca)
nested_resource_name = toscautils.get_nested_resources_name(tosca)
@ -319,7 +322,8 @@ class TOSCAToHOT(object):
heat_template_yaml = toscautils.post_process_heat_template(
heat_template_yaml, mgmt_ports, metadata, alarm_resources,
res_tpl, block_storage_details, self.unsupported_props)
res_tpl, block_storage_details, self.unsupported_props,
unique_id=unique_id)
self.heat_template_yaml = heat_template_yaml
self.monitoring_dict = monitoring_dict

8
tacker/vnfm/plugin.py
View File

@ -812,14 +812,8 @@ class VNFMPlugin(vnfm_db.VNFMPluginDb, VNFMMgmtMixin):
return trigger
def _handle_vnf_monitoring(self, context, trigger):
vnf_dict = trigger['vnf']
if trigger['action_name'] in constants.DEFAULT_ALARM_ACTIONS:
action = trigger['action_name']
LOG.debug('vnf for monitoring: %s', vnf_dict)
self._vnf_action.invoke(
action, 'execute_action', plugin=self, context=context,
vnf_dict=vnf_dict, args={})
vnf_dict = trigger['vnf']
# Multiple actions support
if trigger.get('policy_actions'):
policy_actions = trigger['policy_actions']