I'm just wondering if I can get a 'vmstat 1 10' or such with a high load alert, because on most of the systems I administrate the vmstat commands first line of output is a cache.
illustrated..:
[root@holiday csf]# date && vmstat 1 1
Fri Jun 26 11:51:56 NZST 2009
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 398580 1009028 28476 265884 16 13 94 46 2 1 5 0 92 3
[root@holiday csf]# date && vmstat 1 1
Fri Jun 26 11:52:07 NZST 2009
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 398580 1007116 28500 265980 16 13 94 46 2 1 5 0 92 3
[root@holiday csf]# date && vmstat 1 1
Fri Jun 26 11:52:37 NZST 2009
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
1 0 398580 1007444 28560 266580 16 13 94 46 2 1 5 0 92 3
[root@holiday csf]# date && vmstat 1 5
Fri Jun 26 11:52:47 NZST 2009
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 398580 1005828 28584 266920 16 13 94 46 2 1 5 0 92 3
0 0 398580 1006052 28592 266912 0 0 12 96 1035 184 7 1 91 1
0 0 398580 1006500 28592 266912 0 0 0 0 1034 90 0 0 100 0
1 0 398580 1007076 28592 266912 0 0 0 0 1012 207 0 0 100 0
1 0 398580 1007028 28592 266912 0 0 0 0 1036 386 5 0 95 0
As you can see from the above, swap/io/system/cpu values aren't truly reflected with a 'vmstat 1 1'
ideas?
thanks!