TRAILHEAD safety checklist¶

TRAILHEAD runs against live clusters. Mistakes can disrupt workloads, corrupt storage, or trigger unexpected failovers. Run through this checklist before every cycle.

Pre-run checklist¶

Cluster state¶

• Cluster is not hosting production workloads during the cycle window
• All cluster nodes are online (Get-ClusterNode | Where-Object State -ne 'Up' returns nothing)
• Storage pool is Healthy (Get-StoragePool | Select-Object HealthStatus, OperationalStatus)
• No pending storage repairs (Get-StorageJob returns nothing or only scheduled tasks)
• No active cluster events with severity Error or Critical in the last 30 minutes

Tool version¶

• The tool version being tested is pinned (commit SHA, PSGallery version, or tagged release)
• The tool version matches the cycle scenario — not HEAD/main unless intentional
• Any -WhatIf flags are confirmed appropriate for the scenario

Evidence capture¶

• Evidence log path exists and is writable
• Sufficient disk space (>500 MB free) on the log destination
• Start-TrailheadRun.ps1 has been invoked to initialise the run log

Communication¶

• Other team members know a TRAILHEAD cycle is in progress on this cluster
• A rollback plan is documented if the tool under test makes cluster changes

During the run¶

• Do not run other tests or management operations on the cluster simultaneously
• If a step fails and you don't understand why — stop the cycle, investigate, then restart
• All failures must be logged with Write-THFail before continuing or aborting

Post-run checklist¶

• Close-THRun was called to write the summary footer to the log
• All evidence files are saved in the run's evidence directory
• Cluster state is verified healthy after the run
• Log directory is committed/pushed to the consumer repo (or retained per team procedure)
• If failures occurred: a GitHub issue is open with the cycle log attached or linked