| Doug Freyburger 2007-05-10, 1:28 pm |
| Folks,
One of my Solaris 10 boxes is giving unreasonable numbers in vmstat.
SunOS $HOST 5.10 Generic_118833-17 sun4u sparc SUNW,Sun-Fire-V490
# vmstat 10
kthr memory page disk faults
cpu
r b w swap free re mf pi po fr de sr s3 sd sd sd in sy cs
us sy id
0 0 0 14867184 13606744 156 165 1572316795403 15 15 0 3 0 40 40 20
2592 1579 1926 1 2 96
0 0 0 14553144 13399904 120 604 3431684232 2 2 0 0 0 12 11 10 680
1492 687 1 2 97
0 0 0 14545112 13393008 166 508 0 922 916 0 0 0 2 2 95 1603 13299
6885 12 7 81
0 0 0 14548384 13399264 10 0 3434256708 0 0 0 0 0 3 3 13 1202 586 884
0 1 99
0 0 0 14548544 13398496 19 8 0 0 0 0 0 0 6 6 14 1246 698 933
0 2 98
0 0 0 14549784 13397552 10 0 0 0 0 0 0 0 0 0 11 1069 527 809
0 1 99
^C
Ignoring the first line as usual the page-in line bounces between
small
numbers or zero and unreasonably large numbers in the billions.
There is no way billions of pages of data are being read into memory
even with a large file being mapped to virtual memory and reading it
as fast as the IO system can deliver.
Not ignoring the first line, the page-in number of the average since
boot is even more unreasonable. This system runs an Oracle
production database and it's busy (90+% idle CPU is less common
than in the slice I just pasted).
I take it this is a reporting error of some sort? Where else should I
look to see if any action is needed.
Why I ask - This host has sieved twice in 6 months to the point a
break on the console didn't help and had to be power cycled. Last
night was the second time. This vmstat output is the most unusual
thing on the system while it's running.
|