Skip to content

Allow fuzzing janitor to delete Batch jobs - #355

Merged
markcor merged 1 commit into
mozilla-platform-ops:masterfrom
pyoor:fuzzing-janitor-job-delete
Nov 3, 2026
Merged

markcor merged 1 commit into
mozilla-platform-ops:masterfrom
pyoor:fuzzing-janitor-job-delete

Conversation

@pyoor

@pyoor pyoor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

Adds job-delete permission to the fuzzing Batch janitor role so cleanup can force-delete expired jobs before deleting their pools, avoiding stalled cleanup when a node is unreachable.

@markcor

markcor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

@pyoor, any reason I shouldn't apply this change?

@pyoor

pyoor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor Author

@markcor Nope, it's ready to land. Thanks!

@markcor
markcor merged commit 744e704 into mozilla-platform-ops:master Nov 3, 2026
1 check passed
@markcor

markcor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

@pyoor The Terraform plan detects that fuzzing-agent-subnet was deleted outside Terraform. Applying the full plan as-is would:

  • Add Microsoft.Batch/batchAccounts/jobs/delete to the Fuzzing Batch harness janitor role, as intended by this PR.
  • Recreate fuzzing-agent-subnet (10.1.3.0/24) in the clauditor-run-vnet virtual network, resource group clauditor, with default outbound access disabled.
  • Recreate its association with fuzzing-agent-subnet-nsg.
  • Recreate its association with fuzzing-agent-subnet-routes.
  • Destroy and recreate the subnet_join role assignment on that subnet.

The plan totals 4 to add, 1 to change, 1 to destroy. No apply has been run as part of this work.

Do you want that subnet and its associations/role assignment restored, or was its deletion intentional and we should keep this apply limited to the janitor permission change?

@pyoor

pyoor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor Author

In response to the Terraform drift: please restore fuzzing-agent-subnet, its NSG and route-table associations, and the subnet-join role assignment, then apply the janitor job-delete permission.

The subnet’s removal was unintended. Azure history shows a November 2 operator VNet update removed it. The Windows isolation branch already includes a fix to preserve additional subnets during reconciliation.

@markcor

markcor commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

Changes appplied.

@pyoor
pyoor deleted the fuzzing-janitor-job-delete branch November 3, 2026 20:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants