Fix a bug that causes piped crashes while rolling back

[nghialv]

What this PR does / why we need it:

Scenario:

• A deployment is triggered
• The first stage (e.g. K8S_SYNC) is cancelled from web UI while ensurePreparing is running
• So prepareOnce is marked as executed but s.deploymentConfig = cfg is still not configured
• The rollback stage starts running
• Because the prepareOnce was marked as executed so s.deploymentConfig is still nil
• That causes a panic when accessing s.deploymentConfig in rollback execution

Which issue(s) this PR fixes:

Fixes #

Does this PR introduce a user-facing change?:

NONE

/cc @nakabonne

[pipecd-bot]

Some issues were detected while linting go source files in your changes.

[pipecd-bot]

Code coverage for golang is 21.50%. This pull request decreases coverage by -0.01%.

File	Function	Base	Head	Diff
pkg/app/piped/controller/scheduler.go	scheduler.ensurePreparing	0.00%	0.00%	+0.00%

[pipecd-bot]

Code coverage for golang is 21.50%. This pull request decreases coverage by -0.01%.

File	Function	Base	Head	Diff
pkg/app/piped/controller/scheduler.go	scheduler.ensurePreparing	0.00%	0.00%	+0.00%

[nakabonne]

Other than some nits, looks good to me.

[nakabonne]

/approve

[pipecd-bot]

This pull request is APPROVED by nakabonne.

Approvers can cancel the approval by writing /approve cancel in a comment. Any additional commits also will change this pull request to be not-approved.

[pipecd-bot]

failed to handle template rule examples

Details

Error: unabled to push rendered files to destination