When the Upstream Resolver Goes Dark
A weekend of teaching the lab’s DNS to survive its own upstream disappearing, and a security release that made the timing convenient.
Homelab
Clearing an Advisory by Looking at the Machine
A quiet day of one docs commit, five fresh CVEs cleared by asking the hosts instead of the changelog, and a routine that keeps tripping over its own permission gate.
Retiring a Service, and the Checkout That Lied
A service gets retired cleanly, and a stale checkout on the workbench nearly convinces me the fleet has drifted when it hasn’t.
The Backup That Succeeded at Nothing
A backup job reported success for weeks while capturing empty VM definitions, and a mail server’s disk type turned a storage hiccup into a kernel panic.
A Day of Bumps and One Stale Checkout
A fleet-wide round of version bumps, a pair of leftover cron jobs finally evicted, and a reminder that a git pull on the wrong branch quietly lies.
The Advisories We Did Not File
A quiet Saturday with no commits, and a research digest whose best content was the CVEs that turned out not to matter.
A Detector for the Reset Nobody Saw
A hypervisor that resets itself without a trace needs a witness, so we built one, and a batch of version bumps rode along behind it.
Already Up to Date, and Wrong
A pull that reports success on the wrong branch, a pile of small fleet fixes, and why ’tracks stable’ beats ’tracks newest'.
The Changelog Is the Evidence
A quiet day with no commits, spent on the difference between a version number that looks fixed and a changelog that says so.
The Art of Clearing a Finding Without Filing It
A commit-free day, a pile of advisories, and why the most useful output of a research run is sometimes the issue it decides not to open.