SUMMARY: load average wrong?
2007-12-24 22:06:00
The overwhelming response was that you need to divide the load average by
the number of CPUs to make a comparison. A 12 processor box with a load
average of 5 is about as busy as a 1 processor box with a load average of
5/12 or 0.42.
To me that doesn't match the definition of load average given in uptime(1)
which says it is "the average number of jobs in the run queue over the last
1, 5 and 15 minutes." If there were an average of 5 jobs in the run queue,
wouldn't vmstat show 5, or at least some non-zero value, in the run queue
column? I am going to guess that the difference is that the "load average"
computation used by w, uptime, and top includes processes that are
*currently* running, where vmstat only reports on processes that are
*waiting* to run. In my stats, vmstat shows 0 in the run queue column -- no
processes waiting to run. But looking at the idle column seems to support
that about half of the 12 CPUs are busy at any given instant, i.e. there are
about 6 running processes at any given instant. So if the load average
includes these running processes, that would support a load average of about
6.
Thanks again for all of your responses!
Doug
-----Original Message-----
From: Granzow, Doug (NCI)
Sent: Wednesday, March 06, 2002 3:25 PM
To: 'codeprof at codeprof.com'
Subject: load average wrong?
I suspect the load average on a Sunfire 6800 is wrong. It is reporting load
averages around 5.00 even though all 12 of the CPUs are partially idle.
vmstat shows r = 0. What could be causing the load average to be high, or
how can I get an accurate report of the load average? (w, uptime, and top
all report the same load average)
bash-2.03# uname -a
SunOS hostname 5.8 Generic_108528-10 sun4u sparc SUNW,Sun-Fire
bash-2.03# uptime
3:17pm up 21 day(s), 1:31, 4 users, load average: 5.18, 5.01, 5.11
bash-2.03# vmstat 5
procs memory page disk faults cpu
r b w swap free re mf pi po fr de sr m1 m1 m1 m2 in sy cs us sy
id
23 0 0 21384600 10314952 0 0 0 0 0 0 0 2 1 1 0 4294967196 0 0 -59 -2
-43
0 0 0 21508048 10822144 125 76 979 0 0 0 0 0 0 0 0 4586 249126 2693 45 6
49
0 0 0 21507664 10817944 119 15 947 0 0 0 0 0 0 0 0 4573 250649 2655 45 6
50
0 0 0 21506864 10813520 119 49 960 0 0 0 0 0 0 0 0 4575 249904 2671 44 6
50
0 0 0 21505792 10809008 109 22 854 0 0 0 0 0 0 0 0 4558 249775 2623 44 8
48
0 0 0 21510992 10811360 112 552 816 0 0 0 0 0 0 0 0 4470 250176 2397 43 6
51
0 0 0 21511248 10808792 71 29 596 0 0 0 0 0 0 0 0 4347 250057 2008 44 6
50
0 0 0 21511240 10806576 69 15 596 0 0 0 0 0 0 0 0 4338 252180 1988 43 6
50
^C
bash-2.03# mpstat 5
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
12 34 0 1242 614 499 51 25 10 4 0 426 57 2 0 42
13 35 0 190 325 292 52 26 10 5 0 596 57 2 0 41
14 36 0 184 325 295 51 26 10 5 0 661 57 2 0 41
15 35 0 190 328 297 51 26 10 5 0 693 57 2 0 41
16 35 0 178 325 290 51 26 10 5 0 552 57 2 0 42
17 35 0 186 325 288 51 26 10 5 0 534 57 2 0 42
18 35 0 181 325 291 52 26 10 5 0 696 57 2 0 41
19 35 0 187 325 290 53 26 10 5 0 568 57 2 0 41
20 35 0 187 325 288 54 26 10 5 0 516 57 2 0 41
21 34 0 197 353 315 51 25 10 6 0 540 57 2 0 41
22 35 0 197 349 313 51 26 10 5 0 673 57 2 0 42
23 35 0 180 325 286 52 26 10 4 0 554 57 2 0 42
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
12 0 15 3860 611 500 393 15 25 28 0 220 39 2 0 59
13 0 3 132 309 298 52 9 8 8 0 161359 47 34 0
19
14 0 2 95 310 298 55 10 12 9 0 35515 65 12 0 23
15 26 6 217 312 297 142 12 16 20 0 51534 54 11 0 35
16 1 5 243 310 291 92 10 15 22 0 233 58 1 0 40
17 16 15 546 314 293 325 13 21 34 0 312 38 2 0 60
18 0 9 453 313 294 308 13 20 29 0 244 46 2 0 52
19 0 6 179 312 294 108 12 19 15 0 209 51 1 0 47
20 0 13 622 312 293 417 12 19 49 0 439 29 3 0 68
21 0 15 634 924 911 313 12 16 116 0 366 24 2 0 75
22 0 10 473 319 300 263 13 20 33 0 248 39 1 0 60
23 0 13 345 314 294 177 14 23 29 0 329 41 1 0 58
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
12 0 7 3734 606 500 130 10 16 31 0 33540 41 8 0 52
13 13 12 348 312 292 314 12 20 22 0 197 36 3 0 61
14 0 4 134 309 294 74 9 16 11 0 1126 63 1 0 36
15 0 5 130 312 295 75 11 13 18 0 69333 56 14 0 30
16 0 7 262 310 289 100 11 15 24 0 42551 44 11 0 45
17 0 10 247 307 291 113 7 15 18 0 192 34 8 0 58
18 0 6 173 311 291 81 11 14 13 0 18969 56 5 0 39
19 0 6 167 309 289 82 9 15 15 0 974 60 2 0 38
20 1 8 561 310 289 138 10 17 42 0 381 49 2 0 49
21 0 17 813 874 862 1095 30 23 64 0 1206 12 3 0 85
22 0 9 319 317 299 120 11 15 20 0 69102 44 15 0 42
23 0 7 458 314 291 173 14 20 41 0 12015 39 5 0 56
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
12 0 16 3964 611 500 224 15 21 48 0 1675 30 2 0 68
13 0 10 289 309 295 99 9 16 22 0 736 47 2 0 51
14 13 6 154 312 297 91 12 19 12 0 204 59 2 0 39
15 0 6 150 313 296 93 12 16 12 0 42861 63 8 0 29
16 0 10 426 311 294 128 11 19 29 0 5135 47 2 0 51
17 0 6 156 307 293 78 7 13 11 0 137 46 4 0 50
18 0 4 236 310 294 95 10 15 18 0 5281 56 1 0 43
19 0 12 495 315 294 435 15 23 32 0 314 36 2 0 62
20 0 13 502 310 293 370 10 17 40 0 530 39 2 0 59
21 0 16 869 922 911 805 21 17 89 0 565 15 2 0 83
22 0 8 328 318 302 177 11 12 11 0 141415 44 27 0
29
23 0 5 211 310 295 82 10 13 13 0 52874 51 13 0 36
^C
bash-2.03# uptime
3:18pm up 21 day(s), 1:32, 4 users, load average: 5.62, 5.18, 5.16
Will summarize. Thanks in advance,
Doug
_______________________________________________
codeprof mailing list
codeprof at codeprof.com
http://www.codeprof.com/execute/ask/?codeinfoid=11602
Comments
Got something to say?
You must be logged in to post a comment.
