On 25/02/10 12:31 PM, Bruno Damour wrote:
Hello,
On osol-b133 noticed that the first call to vmstat always yields a
significantly different figure for freemem than the subsequent calls :
# vmstat 2
kthr memory page disk faults cpu
r b w swap free re mf pi po fr de sr cd s0 -- -- in sy cs us sy id
0 0 23 199424 60264 9 61 9 14 29 0 749 8 -1 0 0 679 1652 1387 2 1 97
0 0 48 192448 80700 96 121 0 0 0 0 0 0 0 0 0 537 933 446 2 1 98
0 0 48 192368 81176 0 0 0 0 0 0 0 0 0 0 0 476 894 415 1 0 99
0 0 48 192368 81176 0 0 0 0 0 0 0 0 0 0 0 478 943 430 0 1 99
Furthermore, if I let one "vmstat 2" instance running in a shell, and launch a
new one in another, the first call yields a different value than the one reported by the
running one.
I cannot but think that the value on the first line must be unaccurate, which would mean
that the output of "vmstat" (without the time loop) is unaccurate.
I do not see this behaviour in my linux vm guest (gentoo), which reports
consistents values even in the first call.
Is this behaviour expected ? Is it the same on other opensolaris builds, on
SXCE, on solaris 10 ?
Is there any obvious explanation I'm missing ?
First line is a summary of the virtual memory activity since the
system was booted. This behaviour is same in Solaris 10 and OpenSolaris
builds too afaik.
Hemantha
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org