...
| Note |
|---|
We have a Windows Bamboo Agent and a Linux one. There are some small differences in how the jobs need to be setup depending on which OS the job is to run on. Where appropriate this will be indicated below. |
Step 1
Log in to Bamboo and Click the "Create Plan" button.
...
"Bitbucket" is the name for bitbucket.renalregistry.nhs.uk and you can then select from any of the repositories that you have access to.
Step 2
Ignore this screen for now and click "Create". Make sure that the plan is not enabled otherwise it will try to run and fail straight away.
Step 3 - Repositories
Go to the Repositories Screen and link in the "housekeeping" repository. This contains the build scripts that your job will use. Your project repository should be the default.
| Warning |
|---|
Be careful if making changes to the housekeeping scripts as they will be automatically checked out and run by all the other jobs. |
Step 4 - Triggers
This menu can be used to configure additional ways in which your job can be triggered - a common one would be linking the projects that your project depends on so it is rebuilt after they are updated.
This can be left as-is for a simple job.
Step 5 - Branches
This menu controls whether your job also applies to any branches of the default repository.
Typically this would be set to "When new branch in repository is created" so that when you are developing a change it is automatically tested. Bitbucket will automatically tidy up files that relate to branches that have become inactive.
Step 6 - Dependencies / Permissions
Leave all these as they are.
Step 7 Notifications
Add a Notification on the Event "Failed Builds and First Successful" and set it to notify the Responsible User(s). If the build breaks this will e-mail those who had last comitted to the repository until the build is fixed.
Step 8 - Variables
Leave this as it is.
(TODO: Check if we have any examples where we use this...)
Step 9 - Miscellaneous / Audit Log
Leave these as they are.
Step 10 - Job(s)
Select the Default Job from the left hand side of the page. All our build plans (so far) only use a single job.
Step 11 - Source Code Checkout
This step checks out your code to the build directory of the server, along with the housekeeping scripts that will be used by other steps.
...
- There should be a Task called "Source Code Checkout". Click on this and see that it is set to checkout the Default Repository which should be your project.
- Click on the "Add Repository" button at the bottom of the screen and choose the housekeeping repository. Enter housekeeping as the checkout folder.
- Click Save.
Step 12 - Script Task (Tox)
This step runs Tox which performs a number of tasks such as running unit tests and returning their output in a standard way. (TODO: Link to Tox documentation once written...)
...
- Click on the "Add Task" button and search for a "Script" task. Click OK.
- In the Task Description enter "Run Tox".
- LINUX: As this is a Linux task put "Shell" as the Interpreter.
- LINUX: For the script location Select "File" and put "housekeeping/scripts/linux/runtests.sh" as the script location.
- WINDOWS: TODO
- Click Save
Step 13 - Script Task (Build Wheel)
This is only required if you wish to package your code as a module to be distributed via Pip. This isn't needed for all repositories, such as those which contain a collection of scripts.
...
- Click on the "Add Task" button and search for a "Script" task. Click "OK".
- In the Task Description enter "Build Wheel"
- LINUX: Again, as this is a Linux task put "Shell" as the Interpreter.
- LINUX: For the script location Select "File" and put "housekeeping/scripts/linux/build_wheel.sh" as the script location.
- WINDOWS: TODO
- In the Argument field put "dev my_package", where my_package is the name of the package to be built. Dev is the index to which it will be published.
- Click Save
Step 14 - Junit Parser Task
This step takes the XML output produced by any tests that have run and parses it into a format that Bamboo can understand.
- Click on the "Add Task" button and search for the "Junit Parser" task. Click OK.
- In the Task Description enter "Parse Reports"
- The custom results directory should already be set to "**/test-reports/*.xml" - if so leave as-is.
- Click Save
- Drag the "Junit Parser" Task down to the "Final Tasks" section. This is so that it runs even if there has been a problem with the previous steps.
Step 15 - Requirements
- Click on the "Requirements" Tab.
- As this is a Linux job choose "Linux Python3.6" and "exists", then "Add". This should change the "Number of Agents this job can build on" from 2 to 1 as you have restricted it to running on the Linux server.
Step 16 - Artifacts
Artifacts are files that have been produced as part of the build process that you wish to make available via the Bamboo website. This may include code coverage reports or zipped builds of the code.
The locations will vary but if you have used the Build Wheel script above you should find a wheel file in a dist subdirectory of wherever your code was checked out to.
Step 17 - Miscellaneous
You should be able to leave this as-is. Setting it to clean the build directory is sometimes useful if you are having issues with your Python virtualenv containing incorrect package versions but it will noticably slow down the build process.