We have provided functions to get used and free resources in
class ModelRoot. So strategies can invoke the functions to
get used and free resources.
Change-Id: I3c74d56539ac6c6eb16b0d254a76260bc791567c
Use the general purpose threadpool when building the nova compute
data model. Additionally, adds thorough explanation about theory of
operation.
Updates related test cases to better ensure the correct operation
of add_physical_layer.
Partially Implements: blueprint general-purpose-decision-engine-threadpool
Change-Id: I53ed32a4b2a089b05d1ffede629c9f4c5cb720c8
Implements the singleton general purpose threadpool for the decision
engine and associated tests.
A threadpool is a collection of one or more threads typically called
'workers' to which tasks can be submitted. These submitted tasks will
be scheduled by the threadpool and subsequently executed. How many
tasks will be executed concurrently is managed by the underlying
threadpool and its configuration. In Python the submission of tasks
to a threadpool returns an object called a 'future'. Futures provide
a method to interface with the task being executed that allows to
retrieve information about its state. Such as if it currently is being
executed, if it is waiting on a condition and if it has completed
succesfully. Finally, futures allow to retrieve what has been returned
by the submitted task.
In the case of most OpenStack projects instead of interfacing with native
Python concurrency the futurist library is used. This library provides
very similar interfaces to native concurrency with some extras such as
the wait_for_any method.
For more information about futurist or Python concurrency the following
references can be consulted:
https://docs.python.org/3/library/concurrent.futures.htmlhttps://docs.openstack.org/futurist/latest/reference/index.html#executors
Partially Implements: blueprint general-purpose-decision-engine-threadpool
Change-Id: I94bd9a17290967f011762f2b9c787ee7c46ff930
When querying data from datasource, it's possible to miss some data.
In this case if we throw an exception, Audit will failed because of
the exception. We should remove the exception and give the decision
to the strategy.
Change-Id: I1b0e6b78b3bba4df9ba16e093b3910aab1de922e
Closes-Bug: #1847434
Reason:
When there is a compute node but no virtual machine,
the command 'watcher datamodel list' should display
the information of the compute node instead of return None.
Change-Id: Id5ff7f08ac8a9883af9f0313785b756d813ed5a2
Closes-Bug: #1844948
The default planner can not create actions with right order,
The node_resouce_consolidation strategy needs to use its
own planner.
Partially Implements: blueprint node-resource-consolidation
Depends-on: I586e67f782e2965234826634ba3ff51681af4df8
Change-Id: I05b02905a3335a73b6926966de6331c632842293
This component is responsible for selecting an appropriate Planner based
on predefined property value passed to concrete Strategy.
Change-Id: I86de95886df5d7e9558512569601e9ea3babb0e9
Implements: bp watcher-planner-selector
Co-Authored-By: Canwei Li <li.canwei2@zte.com.cn>
This strategy is used to centralize VMs to as few nodes as possible
by VM migration. User can set a input parameter to decide how to
select the destination node.
Implements: blueprint node-resource-consolidation
Closes-Bug: #1843016
Change-Id: I104c864d532c2092f5dc6f0c8f756ebeae12f09e
Many strategies need get node used or free resources, we define
two new method for the purpose in ModelRoot class.
Change-Id: I8cb41fd560dbac9a78d25bfdba51799533db83c2
1. Add datamodel api and policy_enfoce file.
2. Add related unittest for data_model api and policy.
Partially Implements:blueprint show-datamodel-api
Change-Id: I1654685d8cf04db5dd132d43a8640ddf91893cad
1. Add datamodel list endpoint and rpc process.
2. Add datamodel list parased and return.
3. Add related unittest.
Partially Implements:blueprint show-datamodel-api
Change-Id: I758b7ca2bc3d8d596d3457277744336c6629bc4e
The fields disk and disk_capacity have the same value,
we just need one, so remove disk_capacity field.
Partially Implements: blueprint improve-compute-data-model
Change-Id: If3d385c5e61713bbdc85e22f10cd75e161ff79f0
For Compute node, we can use the new property to calculate
resource(VCPU, memory and disk).
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: I9fe58603692a9850e86a2c36ad7a31c473070100
For Compute node, when calculating resource(VCPU, memory and disk)
capacity, we need to consider reserved resource and allocation ratio.
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: I70257dd5fb342a67a3ffda1055eddc54b8360ca3
For Compute node, we can use the new property to calculate
resource(VCPU, memory and disk).
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: I2bb230b5f5a573fb3045261dfdee73f1a8434e0d
For Compute node, we can use the new property to calculate
resource(VCPU, memory and disk).
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: I4f041ad25353d575c276fce87fe13c5e6705754f
For Compute node, we can use the new property to calculate
resource(VCPU, memory and disk).
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: Id113b4c19792946329e9ff448bfe636cc8eca057
For Compute node, we can use the new property to calculate
resource(VCPU, memory and disk).
Partially Implements: blueprint improve-compute-data-model
Depends-on: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Change-Id: I7872265b2378e5dc37aa2e086ff1f7fb9071db0b
The node resource(vcpu, memory and disk) used infomation need
to change when creating or deleting instances. Now Placement do
not send notifications, so there is not a good way to capture
the change. We remove these fields and leave the process to strategy.
Partially Implements: blueprint improve-compute-data-model
Change-Id: I3f9a3279a26f3df444117d9265e74cca57b38d6e
Resource(VCPU, memory and disk) capacity need to be calculated
through formula: capacity = (total-reserved)*ratio.
Partially Implements: blueprint improve-compute-data-model
Change-Id: I15ca66dd2c3a21c5acfebf6f04fa6601aff7918f
We have some new fields(vcpus_ratio, vcpus_used, ...)
in the Watcher ComputeNode. During the process of updating
data model by notifications, we need to get data from
placement.
Partially Implements: blueprint improve-compute-data-model
Change-Id: I10587e93bb3e7be6af78bb3a50509d82d8228f78
The node.free_disk_gb does not take allocation ratios used
for overcommit into account so this value may be negative.
We do not need this field and plan to set disk to total disk
capacity and then remove disk_capacity.
Partially Implements: blueprint improve-compute-data-model
Change-Id: I72c4490f5a8d0fbd1039f70ff20f07b743b6bb2d
