The document discusses heat optimization in OpenStack, focusing on orchestration of composite cloud applications using both native hot templates and AWS CloudFormation templates. It outlines challenges such as resource management and database handling while promoting convergence within stack management through specific configurations. Additionally, it highlights strategies for script optimization, resource monitoring, and the use of software deployment groups.

1. Heat Optimization
Rico Lin
Chief OpenStack Technologist, inwinSTACK
IRC: ricolin
rico.l@inwinstack.com
@ricolintw

2. Placeholder Footer Copy / BU Logo or Name Goes Here
Why NOT
> 80%
adoption rates
Orchestrates multiple composite cloud applications by using either the
native HOT template format or the AWS CloudFormation template
format, through both an OpenStack-native REST API and a
CloudFormation-compatible Query API.

3. • Was not parallel excution
• Database monster
• Memory monster
• Don’t know how we can use it?
Because!

4. Convergence :
Truly using every power we have
Template Heat-API AMQP
Heat-engine
AMQP
(engine_w
orker)
Worker
DB
B
A
Template Heat-API AMQP
Heat-engine
DB stackA
stackA
stackA
stackA
?

5. • Start Newton, `convergence_engine` default
to `True`
• (CLI)Migrate your old stack
to convergence_stack
convergence_engine = True
Template Heat-API AMQP
Heat-engine
AMQP
(engine_w
orker)
Worker
DB

6. num_engine_workers =
max(4, multiprocessing.cpu_count())
• What your node been used for?
• How many socket/process/core?
How many workers you needs
Worker 1
...
...
Worker N

7. More Database comsume
when we change to Convergence

8. Eager and Lazy Fetch
• What always been query right next to each
other?(Eager fetch it!)
– raw_template always after stack
• What we can took a chance and read if
required(Lazy fetch it!)
– stack_tag not always after stack

9. Contexts Cache（WIP）
• Only apply when `refresh = False`
• Replace database access
Task
Cache
DB

10. (WIP)Batch Schedule Task
• Accept Queue to process tasks in batch
Template Heat-API AMQP
Heat-engineAMQP
(engine_worker)
Worker
DB
TaskTask
Task
Task
Task
Task
Task
Task
Task
Task
Task

11. Optimize Sync Table
C
D
B
A
SyncPoint
leaves
• Resolving back to back retry
– Random retry
• max_wait=conflict_count*0.01
• wait=random(0, max_wait)
– (WIP)Exponential backoff
• wait=random(max_wait)*2^N+jitter

12. (WIP)Resolving Update Cancel
• How to handle with ½ instances created and
½ are not.
• API for handle_%s_cancel

13. Reduce dramatically (Witout WIP)

14. Placeholder Footer Copy / BU Logo or Name Goes Here
Optimize
your script
More operate = More mistake
1 Operation
<20 Min.
> 100 Operations
>100 Min. VS
Stack

15. Stack = Parameters + Resources
parameters:
flavor:
type: string
network:
type: string
resources:
service_server_group:
type: OS::Heat::ResourceGroup
properties:
count: { get_param: num_servers}
resource_def:
type: OS::Nova::Server
properties:
image: { get_param: image_id }
flavor: { get_param: flavor }
key_name: { get_param: key_name }
networks: [{network: {get_param: network} }]

16. Functions

17. Easy Constraint, Easy life
Parameters:
flavor:
type: string
description: Flavor for the server to be
created
constraints:
- custom_constraint: nova.flavor
volume_size:
type: number
description: Size of volume to attach to
instance
default: 2
constraints:
- range: { min: 1, max: 10 }

18. A good way to monitor your
resources

19. Making Good Use of Old Resources
resources:
cinder_volume:
type: OS::Cinder::Volume
properties:
size: { get_param: volume_size }
image: { get_param: image }
service_server_group:
type: OS::Heat::ResourceGroup
properties:
count: { get_param: num_servers }
resource_def:
type: OS::Nova::Server
properties:
image: { get_param: image_id }
flavor: { get_param: flavor }
key_name: { get_param: key_name }
networks: [{network: {get_param: network} }]
Nothing change!

20. Nested Stack -> type: stack.yaml
resources:
cinder_volume_attach:
type: cinder_volume_attach.yaml
properties:
size: { get_param: volume_size }
...
service_server_groups:
type: server_group.yaml
properties:
...
cinder_volume:
type: OS::Cinder::Volume
properties:
size: { get_param: volume_size }
image: { get_param: image }
service_server_group:
type: OS::Heat::ResourceGroup
properties:
count: { get_param: num_servers }
resource_def:
type: OS::Nova::Server
properties:
image: { get_param: image_id }
flavor: { get_param: flavor }
key_name: { get_param: key_name }
networks: [{network: {get_param: network} }]

21. my_res_group
type: OS::Heat::ResourceGroup
properties:
count: 100
resource_def:
type: OS::Nova::Server
…
Resource Group
scale_up_policy:
type: OS::Heat::ScalingPolicy
properties:
auto_scaling_group_id: { get_resource: group }
...
group:
type: OS::Heat::AutoScalingGroup
...
type: OS::Nova::Server
...
outputs:
server_list:
value: [{get_attr: [my_res_group, name]}]

22. Software Config
my_deps:
type: OS::Heat::SoftwareDeploymentGroup
properties:
config: {get_resource: config}
servers: {{get_attr: [servers, name]}}
signal_transport: NO_SIGNAL
config:
type: OS::Heat::SoftwareConfig
properties:
group: ansible
inputs:
- name: cluster-one
outputs:
- name: result
config:
get_file: config-scripts/cluster-ansible.ansible
---
- name: Mysql Cluster -
connection: local
hosts: localhost
tasks:
- name: Hello touch_file
shell: echo {{ foo }} >> /tmp/{{ bar }}
- name: Hello echo
shell: echo "The file /tmp/{{ bar }} contains {{ foo }}"
>> {{ heat_outputs_path }}.result

23. Multi layer services
Another_SoftwareDeploymentGroup
config: {get_resource: another_config}
servers: {{get_attr: [servers, name]}}
signal_transport: NO_SIGNAL
Another_config:
type: OS::Heat::SoftwareConfig
properties:
group: ansible
inputs:
- name: {get_attr: [config, result]}
outputs:
- name: result
config:
get_file: cluster_db-ansible.ansible
...
shell: ... >> {{ heat_outputs_path }}.result

24. Use case
DB DB DB Storage Storage Storage
Load Balancer
Auto Scaling
Web Server Web ServerWeb Server
Stack

25. Heat Optimization
Rico Lin
Chief OpenStack Technologist, inwinSTACK
IRC: ricolin
rico.l@inwinstack.com
@ricolintw
Demo

26. Heat Optimization
Rico Lin
Chief OpenStack Technologist, inwinSTACK
IRC: ricolin
rico.l@inwinstack.com
@ricolintw
Q&A