Rundeck - loop on referenced job - rundeck

Question

I break my head with a problem of referenced job on my workflow. I'm not sur this is possible with Rundeck :
I have a job who call a second. I want run this second for all nodes but only over one server.
With this exemple maybe it's more simple to understand:
Workflow : Select Nodes
Referenced job 1
NodeA > Website www.exempleA.com < restore DB with default value
NodeB > Website www.exempleB.com < restore DB with default value
NodeC > Website www.exempleC.com < restore DB with default value
NodeD > Website www.exempleD.com < restore DB with default value
This run perfectly
Referenced Job 2 : Use Cypress server to test websites. Node filter have only Cypress server.
NodeE > Cypress -url https://${node.name} = NodeA > www.exempleA.com
NodeE > Cypress -url https://${node.name} = NodeB > www.exempleB.com
NodeE > Cypress -url https://${node.name} = NodeC > www.exempleC.com
NodeE > Cypress -url https://${node.name} = NodeD > www.exempleD.com
So I want to make a loop with a referenced job who execute on only one server but for all nodes name.
Someone know if this configuration is possible with Rundeck ?
Thank you for your knowledge.
Erwan

Answer 1

An excellent way to do that is to play with parent job options in two ways: first, against the first child job as a node filter (to dispatch to remote nodes), and second, against the second child job (to create an array and run the Cypress command in a bash loop).
Here is an example to test.
Parent Job. Contains an option that should be used for child jobs.
- defaultTab: nodes
description: ''
executionEnabled: true
id: db051872-7d5f-4506-bd49-17719af9785b
loglevel: INFO
name: ParentJob
nodeFilterEditable: false
options:
- name: nodes
value: node00 node01 node02
plugins:
ExecutionLifecycle: null
scheduleEnabled: true
sequence:
commands:
- jobref:
args: -myfilter ${option.nodes}
group: ''
name: FirstChildJob
nodeStep: 'true'
uuid: f7271fc4-3ccb-41a5-9de4-a12e65093a3d
- jobref:
args: -myarray ${option.nodes}
childNodes: true
group: ''
name: SecondChildJob
nodeStep: 'true'
uuid: 1b8b1d82-a8dc-4949-9245-e973a8c37f5a
keepgoing: false
strategy: sequential
uuid: db051872-7d5f-4506-bd49-17719af9785b
First Child Job. Takes the parent job option and uses that as a job filter, the node filter is an own option called ${option.myfilter}.
- defaultTab: nodes
description: ''
executionEnabled: true
id: f7271fc4-3ccb-41a5-9de4-a12e65093a3d
loglevel: INFO
name: FirstChildJob
nodeFilterEditable: false
nodefilters:
dispatch:
excludePrecedence: true
keepgoing: false
rankOrder: ascending
successOnEmptyNodeFilter: false
threadcount: '1'
filter: ${option.myfilter}
nodesSelectedByDefault: true
options:
- name: myfilter
plugins:
ExecutionLifecycle: null
scheduleEnabled: true
sequence:
commands:
- exec: echo "hi"
keepgoing: false
strategy: node-first
uuid: f7271fc4-3ccb-41a5-9de4-a12e65093a3d
Second Child Job. Contains an inline-script step that takes the parent's job option as an array and runs in a bash loop.
- defaultTab: nodes
description: ''
executionEnabled: true
id: 1b8b1d82-a8dc-4949-9245-e973a8c37f5a
loglevel: INFO
name: SecondChildJob
nodeFilterEditable: false
nodefilters:
dispatch:
excludePrecedence: true
keepgoing: false
rankOrder: ascending
successOnEmptyNodeFilter: false
threadcount: '1'
filter: 'name: node02'
nodesSelectedByDefault: true
options:
- name: myarray
plugins:
ExecutionLifecycle: null
scheduleEnabled: true
sequence:
commands:
- script: "#!/bin/bash\narray=(#option.myarray#)\nfor i in \"${array[#]}\"\ndo\n\
\techo \"execute $i\"\ndone"
keepgoing: false
strategy: node-first
uuid: 1b8b1d82-a8dc-4949-9245-e973a8c37f5a
Here is the loop script (inside the second child job as inline-script):
#!/bin/bash
array=(#option.myarray#)
for i in "${array[#]}"
do
echo "$i"
done
And here you can see the result.