Toad (toad at amphibian.dyndns.org) wrote:
> What conclusions can be drawn? Mainly, that system load and both local mean
> traffic and the load percentage have absolutely nothing to do with each
> other.
I've been saying that for a while now. ;-)
> One thing I have suggested in the past has been to measure
> the actual CPU usage, or the system load, on an architecture specific
> basis (on linux, parse /proc/stat, on other unix, parse the output of
> vmstat, on Windoze, use JNI, etc). I would like to point out here that
> transparent portability to arbitrary future or obscure platforms IS NOT
> AND NEVER HAS BEEN A CORE PROJECT GOAL. I have functional java code to
> parse /proc/stat and determine CPU every few seconds (on linux). I
> believe fish has JNI code to find the CPU usage from the JVM on
> Windoze. Since load balancing is fairly critical to a functioning
> Freenet, I would like to add this code to Fred at least as an option.
> Does anyone agree with me?
I think it's a good idea, and deserves to be tried, at least in the
unstable branch. I currently have shell access to 3 different kinds
of Unix: Linux (several different versions), HP-UX 10.20, and OpenBSD
3.2. Here are the vmstat outputs from these systems, to help write
the parsing code.
Linux (Debian unstable x86):
griffon:~$ vmstat 10
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
1 0 46932 4424 16900 343268 0 0 13 12 2 12 99 0 0 0
4 0 46932 4416 16900 343268 0 0 0 0 636 1222 98 2 0 0
2 0 46932 4420 16900 343012 0 0 0 0 174 629 100 0 0 0
4 0 46932 4380 16900 343016 0 0 0 7 161 603 100 0 0 0
5 1 46932 4672 16900 342636 0 0 0 0 290 1047 99 1 0 0
3 0 46932 4400 16900 342892 0 0 0 0 202 781 100 0 0 0
2 0 46932 4396 16900 342892 0 0 0 4 177 627 100 0 0 0
4 0 46932 4384 16900 342892 0 0 0 0 178 642 100 0 0 0
3 0 46932 4388 16900 342892 0 0 0 0 180 612 100 0 0 0
4 0 46932 4388 16900 342892 0 0 0 0 171 609 100 0 0 0
3 0 46932 4380 16900 342892 0 0 0 4 191 660 100 0 0 0
4 0 46932 4412 16900 342892 0 0 0 0 205 745 100 0 0 0
4 0 46932 4420 16900 342892 0 0 0 0 174 620 100 0 0 0
3 0 46932 4420 16900 342892 0 0 0 4 199 719 100 0 0 0
2 0 46932 4424 16900 342892 0 0 0 0 172 621 100 0 0 0
4 0 46932 4424 16900 342896 0 0 0 0 187 629 100 0 0 0
4 0 46932 4420 16900 342896 0 0 0 0 198 715 100 0 0 0
4 0 46932 4432 16900 342896 0 0 0 9 179 637 100 0 0 0
3 0 46932 4432 16900 342896 0 0 0 0 195 702 100 0 0 0
4 0 46932 4432 16900 342896 0 0 0 0 180 607 100 0 0 0
2 0 46932 4432 16900 342896 0 0 0 12 171 602 100 0 0 0
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
8 0 46932 4428 16900 342896 0 0 0 0 170 595 100 0 0 0
2 0 46932 4444 16900 342896 0 0 0 0 201 718 100 0 0 0
4 0 46932 4444 16900 342896 0 0 0 0 216 752 99 1 0 0
4 0 46932 4444 16900 342896 0 0 0 4 193 698 100 0 0 0
4 0 46932 4424 16900 342916 0 0 0 0 185 694 100 0 0 0
3 0 46932 4464 16900 342876 0 0 0 0 185 654 100 0 0 0
HP-UX 10.20:
imadev:~$ vmstat 10
procs memory page
faults cpu
r b w avm free re at pi po fr de sr in
sy cs us sy id
2 1 0 7061 9827 1 0 0 0 0 0 0 153
10 461 3 0 97
1 0 0 7027 9797 0 0 3 0 0 0 0 147
323 124 0 0 100
1 0 0 7571 9797 0 0 0 0 0 0 0 121
263 116 0 0 100
1 0 0 7418 9797 0 0 0 0 0 0 0 118
291 122 0 0 100
1 0 0 6680 9813 0 0 1 0 0 0 0 130
402 127 0 0 100
1 0 0 7178 9781 0 0 0 0 0 0 0 116
304 121 0 0 100
1 0 0 6973 9781 0 0 0 0 0 0 0 119
266 116 0 0 100
1 0 0 7191 9765 0 0 0 0 0 0 0 142
255 119 0 0 100
1 0 0 7107 9781 0 0 0 0 0 0 0 121
267 116 0 0 100
1 0 0 6271 9781 0 0 0 0 0 0 0 113
279 120 0 0 100
1 0 0 6480 9781 0 0 0 0 0 0 0 116
291 114 0 0 100
1 0 0 7341 9797 0 0 0 0 0 0 0 113
266 121 0 0 100
1 0 0 6809 9797 0 0 0 0 0 0 0 118
256 116 0 0 100
1 0 0 6342 9797 0 0 0 0 0 0 0 142
268 119 0 0 100
1 0 0 6665 9797 0 0 0 0 0 0 0 119
282 116 0 0 100
1 0 0 6889 9797 0 0 0 0 0 0 0 113
295 120 0 0 100
1 0 0 7629 9797 0 0 0 0 0 0 0 124
430 123 0 0 100
1 0 0 7453 9781 0 0 0 0 0 0 0 113
276 121 0 0 100
1 0 0 6797 9781 0 0 0 0 0 0 0 119
269 116 0 0 100
1 0 0 6749 9781 0 0 0 0 0 0 0 143
287 120 0 0 100
OpenBSD 3.2 x86:
$ vmstat 10
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr wd0 cd0 in sy cs us sy id
2 3 0 263480 176908 88 0 0 0 0 0 14 0 36 4294966839 211 85
15 0
2 3 0 263480 177008 204 0 0 0 0 0 8 0 1742 195954 636 67 33 0
2 3 0 263484 177004 7 0 0 0 0 0 8 0 1246 303843 253 59 41 0
2 3 0 263484 177004 7 0 0 0 0 0 4 0 1212 322105 206 62 38 0
1 3 0 263488 176996 9 0 0 0 0 0 6 0 1216 236838 220 68 32 0
4 3 0 263488 177080 257 0 0 0 0 0 16 0 1284 263460 268 61 39 0
1 3 0 263488 177080 193 0 0 0 0 0 5 0 1233 284126 217 61 39 0
2 3 0 263488 177080 206 0 0 0 0 0 6 0 1231 278494 216 60 40 0
1 3 0 263492 177076 8 0 0 0 0 0 4 0 1243 220574 260 66 34 0
1 3 0 263492 177076 9 0 0 0 0 0 4 0 1240 121437 263 80 20 0
1 3 0 263492 177076 10 0 0 0 0 0 6 0 1248 106448 317 85 15 0
1 3 0 264988 175584 172 0 0 0 0 0 25 0 1245 149976 278 74 26 0
1 3 0 265024 175548 196 0 0 0 0 0 6 0 1221 169134 235 75 25 0
2 3 0 265056 175516 209 0 0 0 0 0 6 0 1236 147846 251 79 21 0
1 3 0 265056 175516 10 0 0 0 0 0 6 0 1216 76274 226 88 12 0
1 3 0 265068 175504 9 0 0 0 0 0 3 0 1208 192535 205 76 24 0
1 3 0 265072 175500 11 0 0 0 0 0 6 0 1219 91122 216 85 15 0
2 3 0 265072 175500 10 0 0 0 0 0 5 0 1233 134029 240 82 18 0
3 3 0 265100 175456 246 0 0 0 0 0 4 0 1219 316822 201 61 39 0
2 3 0 265244 175284 218 0 0 0 0 0 8 0 1211 156184 191 75 25 0
2 3 0 265100 175372 47 0 0 0 0 0 9 0 1215 183871 175 69 31 0
2 3 0 265100 175372 3 0 0 0 0 0 4 0 1219 128646 201 83 17 0
procs memory page disks faults cpu
r b w avm fre flt re pi po fr sr wd0 cd0 in sy cs us sy id
2 3 0 265100 175372 4 0 0 0 0 0 3 0 1216 285359 187 70 30 0
1 3 0 265104 175368 6 0 0 0 0 0 5 0 1238 136674 228 82 18 0
2 3 0 265108 175364 195 0 0 0 0 0 4 0 1240 224078 289 77 23 0
1 3 0 265116 175356 260 0 0 0 0 0 7 0 1232 234673 240 73 27 0
1 3 0 265116 175356 10 0 0 0 0 0 6 0 1194 110720 178 86 14 0
Basic observations:
* Headers may be one or two lines, and may or may not be repeated
intermittently during the normal running.
* The first line of numbers is a summary-since-reboot and should be
ignored.
* The fields do not begin in predetermined columns; instead they are
whitespace separated. HOWEVER, sometimes fields may run together
which makes parsing difficult. (Suggestion: if the line being parsed
doesn't have the correct number of fields, discard it.)
--
Greg Wooledge | "Truth belongs to everybody."
greg at wooledge.org | - The Red Hot Chili Peppers
http://wooledge.org/~greg/ |
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20030517/38959da3/attachment.pgp>
