[HPDD-discuss] Profiling or Debugging OST activity

Dear Lustre, How do I profile or debug an OST activity. One of the OSS in our pool is seeing a network throughput of about 2.5-3Gbps consistently for the last 3-4 hours and I see RPC waits for pending writes are piling up and connection to this OST is obviously suffering from clients consistently. My attempt here is to identify what is causing this activity and take any proactive steps if possible. Please advise. Amit