Oracle DBA

These questions focus on architectural depth, troubleshooting acumen, performance tuning, HA/DR integrations, and strategic considerations, aligning with enterprise-level expectations.

1. Explain the internals of Oracle RAC interconnect. How do you troubleshoot interconnect latency or packet loss?

• Expected topics: UDP/IP vs. RDS, kfed, cluvfy comp ocr, oswatcher, diagwait, netstat, clustervision.elog, MTU size mismatches.

2. How does Oracle 19c handle split-brain scenarios? What improvements were made over previous versions?

• Discuss the Voting Disk mechanism, Node eviction process, and I/O fencing using CSSD and Fencing daemon (FMON).
• Mention Oracle 19c’s Disk-based Quorum Voting and Grid Infrastructure Management Repository (GIMR) enhancements.

3. What are the key performance tuning considerations in a RAC environment versus a single-instance environment?

• Emphasis on cache fusion latency, gc buffer busy waits, gc cr block lost, global enqueue, and NUMA-awareness in 19c.

4. How do you monitor and diagnose global cache contention (GCS/GES)?

• Use of gv$ views such as:
  • gv$ges_blocking_enqueue
  • gv$gc_element
  • gv$instance_cache_transfer
• Tools: ADDM for RAC, AWR reports per instance, TFA, Orion, SQLTXPLAIN

5. What is the role of SCAN in Oracle RAC? How would you configure SCAN DNS for high availability?

• Explain Single Client Access Name (SCAN), SCAN listeners, and HA DNS integration.
• How SCAN enables location transparency and load balancing across nodes.

6. What are Oracle 19c-specific RAC enhancements?

• Automatic memory management per instance
• RAC on Extended Clusters improvements
• Server Weight-Based Node Eviction
• UDEV persistent device support via Grid Naming Service (GNS)

7. Explain the use and tuning of Redo Transport in RAC + Data Guard configurations.

• Interaction between local and remote redo log writes
• SRL sizing, network compression, Data Guard broker role in failover automation

8. What is the impact of inter-node communication latency on Cache Fusion? How do you benchmark or simulate it?

• Discuss the use of clufence, orion, iperf, CSS daemon logs, and nettest.
• Influence of interconnect bonding or redundant private interfaces.

9. How would you perform rolling patching in Oracle RAC 19c?

• Use of Fleet Patching and Provisioning (FPP)
• Out-of-place patching, patchmgr, DBMS_ROLLING for Grid Infrastructure
• Ensuring minimum service disruption using Transient Logical Standby or Oracle Goldengate as interim layer (optional design)

10. Explain Oracle RAC One Node. What use cases is it best suited for?

• Discuss consolidation benefits, online instance relocation, license optimization, and eventual migration to full RAC.

11. How do you handle node eviction scenarios in RAC?

• Tools: crsctl, olrcheck, diagcollection.pl, cluvfy comp node
• Event tracing in: /var/opt/oracle/crsd/trace, OSWatcher, odacli

12. What are the challenges of running Oracle RAC in a virtualized or cloud environment?

• Shared disk architecture via iSCSI, NFS, or block volumes
• Network virtualization impact on interconnect throughput
• Oracle Cloud Infrastructure (OCI) native RAC support
• Licensing implications in BYOL (Bring Your Own License) scenarios

13. How do you diagnose and resolve high gc-related waits in RAC?

• Interpretation of waits like:
  • gc buffer busy acquire
  • gc current request
  • gc cr multi block request
• Actions: SQL tuning, object partitioning, data locality, instance affinity settings

14. Describe the use of srvctl vs. crsctl vs. olsnodes. What scenarios demand each?

• srvctl: resource management
• crsctl: clusterware control
• olsnodes: node mapping, CRS status check

15. Explain instance affinity and how services can be optimized in RAC.

• Using srvctl to bind services to preferred instances
• Impact on workload balancing, failover, and session stickiness