openstack/tacker-specs
Code Issues Proposed changes

Fix Spec of "Enhance Placement Process" for Bobcat

Browse Source 
"Enhance Placement Process" has the following spec changes in Bobcat.
* In AZ reselection, select AZ considering Anti-Affinity of
  PlacementConstraint
* Change Heat-API to get an error message when "stack create/update"
  fails from "Show stack details" to "List resource events"
  considering the use of rollback option

This patch fixes these changes.

Implements: blueprint enhance-placement
Change-Id: Iaa082354cdcefe01f9ae6a3cc22207c3b28eafab
This commit is contained in:
Ai Hamano 2023-06-16 14:44:01 +09:00
parent 28b1e8beca
commit c3db3d4d1a
1 changed files with 15 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
specs/2023.1/enhance_placement_process.rst
View File

@ -201,12 +201,12 @@ flow:
3) Availability zone reselection policy 3) Availability zone reselection policy
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Availability zones in error are excluded from the reselection Availability zones in error are excluded from the reselection
candidates, and are reselected preferentially from unselected candidates, and are reselected considering Anti-Affinity of
availability zones. PlacementConstraint.
.. note:: .. note::
Affinity/Anti-Affinity of PlacementConstraint and resource states of Resource states of availability zones are not considered during
availability zones are not considered during reselection. reselection.
The availability zone in error can be identified in the following way. The availability zone in error can be identified in the following way.
@ -261,7 +261,7 @@ VDU1-0/VDU1-1/VDU2-0/VDU2-1 are deployed
.. note:: .. note::
The above is an example, and the reselection target is randomly The above is an example, and the reselection target is randomly
selected from unselected availability zones. selected from availability zone candidates.
+ Heal(VDU1-1/VDU2-0) + Heal(VDU1-1/VDU2-0)
@ -346,8 +346,9 @@ VDU1-0/VDU1-1/VDU2-0/VDU2-1 are deployed
+ VDU2-3: AZ-1 + VDU2-3: AZ-1
.. note:: .. note::
If there are no unselected availability zones left, randomly select If there are no availability zone candidate considering
a reselection target from the selected availability zones. Anti-Affinity left, randomly select a reselection target from other
than the failed availability zones.
In this case, Anti-Affinity cannot be satisfied. In this case, Anti-Affinity cannot be satisfied.
@ -355,19 +356,19 @@ VDU1-0/VDU1-1/VDU2-0/VDU2-1 are deployed
4) Detection method of insufficient resource error 4) Detection method of insufficient resource error
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When "stack create/update" fails, it is detected from "Show stack details" When "stack create/update" fails, it is detected from "List resource
[#Heat-API]_ of Heat-API response whether the failure is due to events" [#Heat-API]_ of Heat-API response whether the failure is due to
insufficient resources. insufficient resources.
The error message that indicates insufficient resources is extracted The error message that indicates insufficient resources is extracted
from the parameter "stack_status_reason" in the response. from the parameter "resource_status_reason" in the response.
.. note:: .. note::
In the case of insufficient resources, the error occurs after "stack In the case of insufficient resources, the error occurs after "stack
create/update" returns an acceptance response, so the "Show stack create/update" returns an acceptance response, so the "List resource
details" response is used to detect the cause. events" response is used to detect the cause.
The following is an example of an error message stored in The following is an example of an error message stored in
"stack_status_reason" when resources are insufficient. "resource_status_reason" when resources are insufficient.
+ ex1) Set the flavor defined in “OS::Nova::Server” to a large value + ex1) Set the flavor defined in “OS::Nova::Server” to a large value
that cannot be deployed (not enough storage/not enough vcpu/not enough that cannot be deployed (not enough storage/not enough vcpu/not enough
@ -536,4 +537,4 @@ References
.. [#Keystone-API_services] https://docs.openstack.org/api-ref/identity/v3/?expanded=list-services-detail#list-services .. [#Keystone-API_services] https://docs.openstack.org/api-ref/identity/v3/?expanded=list-services-detail#list-services
.. [#Heat-API] https://docs.openstack.org/api-ref/orchestration/v1/index.html?expanded=show-stack-details-detail#show-stack-details .. [#Heat-API] https://docs.openstack.org/api-ref/orchestration/v1/index.html#list-resource-events