Azure Devops YAML Pipelines - While loop? - azure

Question

Is there such a thing in Azure Release Pipelines (YAML) as a while loop? My use case is that I am using runOnce strategy to deploy artifacts to a clean environment, deploy a clients data and before I move onto the next client I need to run a query to ensure all the processing is finished and health checks are done. All checks can be done via SQL scripts into an Azure SQL Database and eventually I need to compare the results and task timings against an expected set.
i.e Does processing the client data across branches yield the expected results and timings.
Might be a square peg round hole so happy to use a different approach if there is an easier way.
- deployment : Install
pool:
vmImage: ubuntu-latest
environment:
name: 'Test_Env'
resourceType: VirtualMachine
strategy:
runOnce:
deploy:
steps:
# Remove and re-create blank database on the Elastic Pool.
- task: SqlAzureDacpacDeployment#1
displayName: Drop DB
inputs:
azureSubscription: 'Azure'
AuthenticationType: 'server'
ServerName: '$(DB_SERVER)'
DatabaseName: 'master'
SqlUsername: '$(DB_USERNAME)'
SqlPassword: '$(DB_PASSWORD)'
deployType: 'InlineSqlTask'
SqlInline: |
IF EXISTS (SELECT name FROM master.sys.databases WHERE name = N'$(DB_DATABASE)') DROP DATABASE [$(DB_DATABASE)]
IpDetectionMethod: 'AutoDetect'
- task: SqlAzureDacpacDeployment#1
displayName: Create DB
inputs:
azureSubscription: 'Azure'
AuthenticationType: 'server'
ServerName: '$(DB_SERVER)'
DatabaseName: 'master'
SqlUsername: '$(DB_USERNAME)'
SqlPassword: '$(DB_PASSWORD)'
deployType: 'InlineSqlTask'
SqlInline: |
CREATE DATABASE $(DB_DATABASE) ( SERVICE_OBJECTIVE = ELASTIC_POOL (name = [SQL_ElasticPool] ));
IpDetectionMethod: 'AutoDetect'
- task: CmdLine#2
displayName: Install Product
inputs:
script: |
start /wait msiexec.exe /i "$(System.ArtifactsDirectory)\installer.msi" client_data= $(client_data) DB_USERNAME=$(DB_USERNAME) DB_PASSWORD=$(DB_PASSWORD)
workingDirectory: $(System.ArtifactsDirectory)
- task: CmdLine#2
displayName: Start Service
inputs:
script: |
sc start $(WIN_SERVICE)
# This is where I would want a while-loop
- task: SqlAzureDacpacDeployment#1
displayName: Check if processing finished
inputs:
azureSubscription: 'Azure'
AuthenticationType: 'server'
ServerName: '$(DB_SERVER)'
DatabaseName: '$(DB_DATABASE)'
SqlUsername: '$(DB_USERNAME)'
SqlPassword: '$(DB_PASSWORD)'
deployType: 'InlineSqlTask'
SqlInline: |
select 1 from eventlog if complete = 0
IpDetectionMethod: 'AutoDetect'

Answer 1

Azure Devops YAML Pipelines - While loop?
You could create a template which will have a set of Check if processing finished task, and pass loop time as parameters across during your build, like:
- template: CheckProcessingFinished.yaml
parameters:
param: ["1","2","3"]
CheckProcessingFinished.yaml:
parameters:
param : []
steps:
- ${{each Looptimes in parameters.param}}:
- task: SqlAzureDacpacDeployment#1
displayName: Check if processing finished
inputs:
azureSubscription: 'Azure'
AuthenticationType: 'server'
ServerName: '$(DB_SERVER)'
DatabaseName: '$(DB_DATABASE)'
SqlUsername: '$(DB_USERNAME)'
SqlPassword: '$(DB_PASSWORD)'
deployType: 'InlineSqlTask'
SqlInline: |
select 1 from eventlog if complete = 0
IpDetectionMethod: 'AutoDetect'
- task: PowerShell#2
inputs:
targetType: 'Sleep 30 seconds'
script: |
Start-Sleep 30
You could check the document Solving the looping problem in Azure DevOps Pipelines for some more details.

Answer 2

In YAML we use each keyword, and it is an equivalent of the for loop .
You cannot run a task until a condition is met , you can run multiple
task either by writing them one by one or by using the each
statement which is equivalent to a for loop.
But you can use a GATE which is a way to control a deployment. They are predominantly used for heath checkups of infrastructure, external approvals for deployment, etc.
The gates can either be at the start of the pipeline or at the end of
the pipeline.
Gates are of many types like the ones which use invokes functions and azure monitors and some gates use rest api too. Also you can create your custom gates too.
Gates will continue to check for a user specified condition until it is met.
So, you can use a azure function gate run the required scripts and return the results in the gate for validation.
Refence:
Azure Pipeline Gate