Hello,
I'm seeing this problem with vmstat, where the first line of output always has the same CPU statistics. For example:
neked@nekedmachine:~$ date && vmstat
Fri Jul 24 06:57:08 EDT 2009
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free buff cache si so bi bo in cs us sy id wa st
0 0 130152 293444 23964 356264 0 0 13 125 0 2 7 7 82 4 0
neked@nekedmachine:~$ date && vmstat
Fri Jul 24 06:57:09 EDT 2009
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free buff cache si so bi bo in cs us sy id wa st
0 0 130152 293444 23964 356264 0 0 13 125 0 2 7 7 82 4 0
neked@nekedmachine:~$ date && vmstat
Fri Jul 24 06:57:11 EDT 2009
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free buff cache si so bi bo in cs us sy id wa st
2 0 130152 293352 23964 356264 0 0 13 125 0 2 7 7 82 4 0
neked@nekedmachine:~$ date && vmstat
Fri Jul 24 06:57:13 EDT 2009
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free buff cache si so bi bo in cs us sy id wa st
0 0 130152 293368 23968 356276 0 0 13 125 0 2 7 7 82 4 0
As you can see, CPU statistics are always context switch=0, user=7, system=7, idle=82, I/O wait = 4.
You could see this problem (on my machine which is CentOS 5, and on another machine which Debian Lenny) by typing:
watch -n 1 vmstat
However, the problem goes away if you let vmstat update itself at intervals, i.e.:
neked@nekedmachine:~$ vmstat -n 1
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------
r b swpd free buff cache si so bi bo in cs us sy id wa st
0 0 130152 284248 25264 357732 0 0 13 125 0 2 7 7 82 4 0
0 0 130152 284248 25264 357732 0 0 0 0 1010 304 1 8 91 0 0
0 0 130152 284248 25264 357768 0 0 0 0 1030 233 0 1 99 0 0
0 0 130152 284248 25264 357768 0 0 0 0 1022 235 0 1 99 0 0
0 0 130152 284248 25264 357772 0 0 0 12 1024 237 0 2 98 0 0
0 0 130152 284248 25276 357760 0 0 0 744 1024 217 1 2 97 0 0
But you can still see that the first line of vmstat's output above shows the same symptoms.
What explains this? How do I avoid it?
---------- Post updated at 07:32 AM ---------- Previous update was at 07:13 AM ----------
Uhhh... to answer my own inquiry, I should've RTFM. From vmstat's manpage:
The first report produced gives averages since the last reboot. Addi-
tional reports give information on a sampling period of length delay.
The process and memory reports are instantaneous in either case.
I'll leave this post up there as a cautionary tale for those who ask before doing proper research.