[SunHELP] problem with pageouts when ram is free

[Alex Avriette]

This is solaris 5.8. I cant quite post the exact output since I'm using the
machine at the moment (five instances of tar that cant be stopped). Here is
the same benchmark however (run with fewer iterations):

 procs     memory            page            disk          faults      cpu
 r b w   swap  free  re  mf pi po fr de sr dd f0 m0 m1   in   sy   cs us sy
id
 0 4 0 1183464 176088 144 859 576 0 0 0  0  0  0  0  0 3027 12165 1870 51 49
0
 1 4 0 1183464 176176 48  0 392 8  8  0  0  0  0  0  0 2735 10986 1790 63 37
0
 2 4 0 1183432 176440 32  0 248 112 112 0 0 0  0  0  0 2394 9936 1721 70 30
0
 0 4 0 1183448 176144 48  0 464 8  8  0  0  1  0  0  0 2848 11421 1887 55 45
0
 1 3 0 1183448 176232 64  0 504 0  0  0  0 12  0  0  0 2809 10452 1734 54 46
0
 0 4 0 1183448 176104 32  0 216 0  0  0  0  0  0  0  0 2824 11878 1894 58 42
0
 0 4 0 1182936 175640 83 298 864 8 8  0  0 51  0  0  0 3067 11952 1904 48 52
0
 2 4 0 1182992 175744 128 859 496 0 0 0  0  0  0  0  0 2780 11595 1769 57 43
0
 1 3 0 1182968 175768 83 410 368 0 0  0  0  0  0  0  0 2709 11392 1786 51 49
0
 0 4 0 1182968 175856 101 205 616 0 0 0  0  1  0  0  1 2957 11409 1902 56 44
0
 0 4 0 1182968 176144 114 0 1088 184 184 0 0 0 0  0  0 3360 11161 1970 53 47
0
 0 5 0 1182864 176304 185 1 1672 0 0  0  0 11  0  0  0 3290 11152 1800 52 47
1
 0 4 0 1183000 176952 385 84 2304 0 0 0  0 89  0  0  0 3217 9659 1607 50 42
8
 1 3 0 1182168 176896 359 140 2136 16 16 0 0 96 0 0  0 3389 8204 1556 50 43
7
 3 4 0 1182072 176728 343 0 2264 0 0  0  0 78  0  0  0 4027 9272 1768 49 48
4
 0 5 0 1182168 176760 321 2 1944 0 0  0  0 60  0  0  0 4050 9844 1788 46 50
5
 0 5 0 1182168 176648 299 1 1848 0 0  0  0 73  0  0  0 4069 9762 1804 50 47
4
 3 5 0 1182072 176352 283 383 1760 0 0 0 0 62  0  0  0 3942 9799 1739 42 54
4
 1 4 0 1182136 176472 286 478 1568 8 8 0 0 46  0  0  0 3958 9999 1728 49 49
3
 0 5 0 1182136 176144 219 1 1624 32 32 0 0 54  0  0  0 4072 9869 1798 55 42
3
 2 3 0 1182136 176056 176 0 920 16 16 0  0 41  0  0  0 4060 9976 1777 53 45
2
 1 4 0 1182136 174808 133 3 720 8  8  0  0 31  0  0  0 3811 9851 1755 55 43
2

That output is from vmstat 1. Its doing much better than it was before,
where I was seeing > 400 po. hopefully outlook wont clober this _too_ much.

thanks,
alex