Preventive Service Planning
Abstract
Here are some commands that can be used to measure the performance of a system and locate any bottlenecks:
who -b When was the system last booted.
Uptime Reports uptime and load averages.
vmstat 5 10 Reports 10 samples every 5 seconds.
vmstat -S Reports swapping rather than paging. Lots of si and so will
indicate memory shortageindicate memory shortage.
Uptime Reports uptime and load averages.
vmstat -s Reports the total events since system boot.
netstat -i 5 Reports the number of packets transmitted and received by
interfaces.
netstat -m Reports streams allocations.
iostat -xc 5 Reports extended and CPU usage. Look for large svc_t
(service time) for busy disks and try to split load across disks.
iostat -xpnce More details.
usr/ucb/ps -aux|head Reports top processes using CPU.
ps -eaL Reports LWP (light weight processes).
mpstat Per processer stats.
pmap -X pid Reports application memory requirement.
sar -u CPU system activity report.
sar -q System activity report for runq. Large runqsize will indicate a
bottleneck.
sar -r System activity report for memory and swap.
sar -k Kernel memory system activity report
Content
UNIX
Was this topic helpful?
Document Information
Modified date:
16 June 2018
UID
swg21303518