What problem does it solve?
This Skill helps diagnose and resolve StarRocks node-level incidents where BE processes crash or run out of memory, and FE nodes stall due to deadlocks, Full GC pauses, or heap leaks.
Core Features & Use Cases
- BE failure triage: Distinguishes Linux OOM kills from internal crashes, then guides core dump review, log inspection, and memory attribution.
- FE contention and GC analysis: Traces deadlocks with LockManager or DeadLockChecker output, and separates Full GC stalls from true FE heap exhaustion.
- Memory leak investigation: Uses mem_tracker, memz, jstack, jmap, and audit-log analysis to identify the workload or module driving memory growth.
- Use case: A production cluster stops responding and queries fail with version errors; this Skill narrows the issue to FE deadlock, captures the right evidence, and points to the recovery path.
Quick Start
Describe the StarRocks node symptom, and I will classify whether it is BE OOM, BE crash, FE deadlock, FE Full GC, or FE heap leak, then outline the exact checks and recovery steps.