ROUND 1 Technical Questions: 1. User complains that his beeline is not working on a server? What all areas you will check and how will you troubleshoot? 2. What is BDR and how it works? 3. Kerberos working process. 4. What is the role of the Zookeeper in the cluster? 5. How will you troubleshoot a long-running job in MR? 6. How will you troubleshoot a long-running job in Spark? 7. User is complaining about a persistent issue that my Spark job driver takes a lot of time in 1st run but runs fine in 2nd re-run and this happens every time. How will you troubleshoot? 8. Do you have experience/hands-on in reviewing JAVA heap dump files? 9. A table is deleted in the cluster and the customer wants to know who deleted it. How will you find it? 10. How & what will you check for high CPU utilization on a server? Non-Technical Questions: 1. How will handle a pissed-off/rude customer on call whose issue is not resolved and is re-occurring. 2. What type of customer as per you is most difficult to handle? 3. Let's say you are in the Sev1 issue and a call with a customer is open for the last 8+ hours. Now your internal engineering guy joins that call and he & the customer lands up into an argument. How will you handle this? ROUND 2 1. Hotspotting in Hbase. 2. Hbase concepts in detail 3. Hbase cluster size 4. What is Hfile? 5. What is rowkey in Hbase? How are they designed? How does it looks like when we look at Hbase master UI? 6. Let's say we have a region in RegionServer, there is a start key and end key. Tell me about it. How it is designed. 7. Hbck tool. What do we do with this? How we troubleshoot with this. 8. Spark architecture. 9. How to troubleshoot a slow-running job in Spark? 10. AWK command usage in Linux Shell Scripting.
