I run servers and keep them running. That means DevOps, Linux, and a fair amount of security work, because a server that's up but owned isn't really up.
- Day to day: deploys, CI pipelines, hardening boxes, and reading logs at hours when nobody should be reading logs.
- Lately: writing small Rust tools for the problems I keep running into on Linux machines.
- Ask me about: Linux servers, CI/CD, locking down SSH, or why the thing that worked yesterday doesn't work today.
whatchanged
Something on your Linux box broke and it worked yesterday. whatchanged --since yesterday puts package upgrades, edited configs, new services and cron changes on one timeline, so you can see what moved. Read-only, no daemon, no network.
whatchanged-action
The same idea for CI: a GitHub Action that shows what a job changed on its runner (packages, /etc, systemd, cron). Useful when a pipeline passes and then the next one mysteriously doesn't.
rotcheck
Disks flip bits quietly, and you usually find out years later when an old photo won't open. rotcheck records a checksum for every file in the folders you care about and tells you what went missing, changed or rotted. Offline, zero dependencies, sha256sum-compatible manifests.
Merged pull requests in other people's projects. This table updates itself every week.
| Project | Merged PRs | Most recent |
|---|---|---|
| AndreaBozzo/dataprof | 9 | #1184 test(python): leave execution_time_ms out of the arrow stream quality comparison |
| CrossGL/crosstl | 1 | #3663 fix(opencl): number program-scope __constant bindings before kernel buffers |
| nikhilb2/simple_invoicing | 1 | #617 fix(backend): sort GSTR-1 Table 13 document numbers naturally |
| skdas20/VaultX | 2 | #15 fix(storage): create the vault and session cache owner-only |
| ludo-technologies/pyscn | 1 | #864 fix: report clone fragment complexity (#860) |
| ansible/receptor | 1 | #1612 fix(receptorctl): ignore leading v when comparing versions |
| storytold/pdfcraft | 1 | #3325 Rail: scroll the panel buttons above the page navigation in short windows (#3282) |
Well, mostly still on it.
- Bash scripts for fun. If I do something twice, the third time is a script. Some of them are even a good idea.
- Hardware. I build, fix and upgrade machines, and I'd rather open the case than open a ticket.
- Gaming. Same rule as production: low latency, high uptime, and blame the network when I lose.
while true; do coffee && deploy || rollback; sleep 8h; done