1. What happens if you can’t SSH to a machine? Structured analysis talks in daily work. People should possess basic experience for sure. Further question: What if ssh is slow?
2. You get low disk warning, what to do? Explain the trouble shooting process and summarize common causes as many as possible. After resolving this issue, people should think how to avoid it happening again or auto detect it in the future. Further question: System load is high, what to do?
3. You open your website and the browsers says it’s not reachable. What do you do? This is quite a typical question, not especially for DevOps. People should have a pretty good understanding for the whole process. And how know how to verify each step properly.
4. How to write initscripts for daemon service? For “service XX start”, configure lockfile and pidfile properly. For “service XXX status”, avoid false positive by light-weight functional test like checking TCP port listening. For “service XX stop”, be prepared for service slow shutdown, instead of kill -9 brutally. Further question: Any alternative to init.d in Linux? initscript, upstart VS Systemd? What drives the change?
5. Given string of a=1, b=2, c=3, how to extract the value of b? The goal here is to look for specifics, such as first-hand scripting experience. People should be familiar popular linux toolkits like awk, sed, grep, bash, etc.
6. What’s the last code you wrote recently? Ideal candidates should be doing effective programming recently and be proud of their achievements. Culture fit and passion for tech matters. Further question: Do you have any side projects? What technical blogs or websites do you follow?
7. To hire new DevOps candidates, what questions you would ask? Totally unexpected! People who love their jobs will stand out.