What problem does it solve?
TiKV fatal process exits including panics, segmentation faults, and restart loops cause immediate cluster downtime and data unavailability, and on-call engineers often waste critical time searching for scattered, unvetted troubleshooting guidance. This skill provides a single, curated reference of known TiKV failure patterns to speed up incident resolution.
Core Features & Use Cases
- Known Failure Pattern Catalog: Pre-documented reference for common TiKV fatal exits including SIGSEGV crashes, missing SST file ingest failures, system clock rollback panics, PITR task-triggered crashes, and coprocessor response overflow errors.
- Actionable Resolution Guidance: Each pattern includes clear log signatures, step-by-step workarounds, and fixed version notes to eliminate ad-hoc research during high-pressure incidents.
- Use Case: When an on-call engineer receives an alert for a TiKV node repeatedly crashing with a FATAL log after pausing a PITR task, this skill immediately surfaces the known tikv#17020 root cause, workarounds to avoid frequent pause/resume operations, and the list of fixed versions to upgrade to.
Quick Start
Use the tikv-panic skill to diagnose a TiKV node that is crashing with a SIGSEGV backtrace on an ARM weak-memory platform.