Leveling up in DevOps: the Art of Bad Shell Scripts
summary
Lots of concrete examples: engineers run root on 5000 instances, one intern's obfuscated one‑liners evolve into a 2,000‑line Ruby script, and ad‑hoc wildcards once broke log collection. Ops builds a unified bashrc in git with shared functions, daily kanban demos, and an IRC chatbot (Ruby) for safe contributions. Tools called out: hostname/cut hacks, parallel‑ssh, strace, Perl piped into SQL. Result: fewer duplicate scripts and smoother oncall, though dangerous one‑offs still pop up.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:07Emilys background
- 2:37Emilys day to day
- 3:30Looking for my peers
- 6:04I adore him
- 6:53Questions
- 10:37Oneliners
- 12:08Bosch History
- 16:35Contacts
- 17:15BashQL
- 17:56Write to disk
- 18:34Estrace
- 19:22Batch RC
- 20:31Source
- 21:28Contents
- 23:01Bad Shell Scripts
- 31:11DevOps
- 32:48QA
related talks
40mBash: the #1 Skill of a DevOps Engineer
41mWhen /bin/sh Attacks: Revisiting "Automate All the Things"
57mGo Away Or I Will Replace You With A Very Little Shell Script
34mDevops or: How I Learned to Stop Worrying and Love the Ops Team
42mWhen /bin/sh Attacks: Revisiting "Automate All the Things"
33mOperations: a developer's guide
40mDistilled Abnormal DevOps Iterations
50mKeynote - More than two decades in shell and ops