Performance Checklists for SREs

TIL

780

System

17

devops

39

huydx viết ngày 03/07/2016

Nhặt được cái checklist cho sysadmin rất hay của Brendan Gregg

Linux Perf Analysis in 60s

• uptime ⟶ load averages
• dmesg -T | tail ⟶ kernel errors
• vmstat 1 ⟶ overall stats by time
• mpstat -P ALL 1 ⟶ CPU balance
• pidstat 1 ⟶ process usage
• iostat -xz 1 ⟶ disk I/O
• free -m ⟶ memory usage
• sar -n DEV 1 ⟶ network I/O
• sar -n TCP,ETCP 1 ⟶ TCP stats
• top ⟶ check overview

Linux Disk Checklist

• iostat –xnz 1 ⟶ any disk I/O? if not, stop looking
• vmstat 1 ⟶ is this swapping? or, high sys time?
• df -h⟶ are file systems nearly full?
• ext4slower 10 ⟶ (zfs*, xfs*, etc.) slow file system I/O?
• bioslower 10 ⟶ if so, check disks
• ext4dist 1 ⟶ check distribution and rate
• biolatency 1 ⟶ if interesting, check disks
• cat /sys/devices/…/ioerr_cnt ⟶ (if available) errors
• smartctl -l error /dev/sda1 ⟶ (if available) errors

Linux Network Checklist

• sar -n DEV,EDEV 1 ⟶ at interface limits? or use nicstat
• sar -n TCP,ETCP 1 ⟶ active/passive load, retransmit rate
• cat /etc/resolv.conf ⟶ it's always DNS
• mpstat -P ALL 1 ⟶ high kernel time? single hot CPU?
• tcpretrans ⟶ what are the retransmits? state?
• tcpconnect ⟶ connecting to anything unexpected?
• tcpaccept ⟶ unexpected workload?
• netstat -rnv ⟶ any inefficient routes?
• check firewall config ⟶ anything blocking/throttling?
• netstat -s ⟶ play 252 metric pickup

Linux CPU Checklist

• uptime ⟶ load averages
• vmstat 1 ⟶ system-wide utilization, run q length
• mpstat -P ALL 1 ⟶ CPU balance
• pidstat 1 ⟶ per-process CPU
• CPU flame graph ⟶ CPU profiling
• CPU subsecond offset heat map ⟶ look for gaps
• perf stat -a -- sleep 10 ⟶ IPC, LLC hit ratio

Tham khảo: http://www.brendangregg.com/blog/2016-05-04/srecon2016-perf-checklists-for-sres.html

Nhân tiện có cái link khá hay : https://software.intel.com/en-us/blogs/2014/05/10/debugging-performance-issues-in-go-programs cho các SRE nào đang dùng golang