Recover deliberately. Find the cause. Fix it permanently.
During a major platform upgrade, a file operation went wrong and the file became corrupted. The immediate priority was not proving who was right. It was protecting production and restoring service.
Stop the damage. Use the known-good backup. Restore the previous working state.
Ian reverted to the previous version using available backups and restored service inside the controlled maintenance window. From the end user's perspective, there was no meaningful customer or business disruption.
With production stable, the investigation could continue safely. Ian worked with the vendor to isolate the corruption, identify the permanent correction and reload the affected file during a subsequent controlled maintenance window.
The incident illustrates a core operating discipline: recovery and root-cause resolution are separate jobs. First protect the operation. Then understand what failed. Then make the permanent correction without gambling with production.
Don't just restore. Resolve.