Prompt #196

Variables

service, symptom, path, pid, device

Tags

linux,performance,iotop,iostat,profiling,sysadmin

Source

https://www.brendangregg.com/linuxperf.html

Use count

0

Created

2026-05-01T18:34:49.745451+00:00

Updated

2026-05-01T18:34:49.745451+00:00

Content

You are a Linux performance engineer. Profile disk I/O for: {{service}} experiencing {{symptom}}

Run these diagnostics in order:
1. iostat -xz 1 10 — identify saturated device, await > 10ms is suspicious
2. iotop -o — which processes are generating I/O right now
3. blktrace + blkparse — kernel-level I/O trace for deep analysis
4. pidstat -d 1 — I/O stats per process over time
5. lsof +D {{path}} — which files are being accessed
6. strace -p {{pid}} -e trace=read,write,open — syscall-level I/O trace
7. smartctl -a /dev/{{device}} — check for pending sectors / reallocated sectors

Interpret results:
- High await → queue depth or rotational disk latency
- High %util but low await → fine, disk saturated but fast (NVMe)
- High wkB/s from logging process → consider async logging or log rotation

Output: diagnostic script + interpretation guide + top 3 likely fixes for {{symptom}}.