Ok. It's taken me a while, but I'm back.
I still have those horrible lags on OpenSolaris. For instance, the lag occurs
each time I launch xload on a host with "nothing else" (well, I mean a clean
window session, no other window and no intensive background or net
application). When I do that, the time xload loads, my host is nearly frozen
(moving the mouse is very slow).
I analyzed my stats. It looks like indeed one of my CPUs gets heavily loaded.
Any idea where this comes from... and how to solve it ?
Reminder: I have an AMD 64 dual core, no my ZFS is not compressed, the problem
occurs with any window manager (fvwm, gnome...), it also happens when I surf on
websites with much flash, or when I launch thunderbird. And I'm running
OpenSolaris 2008.11.
The stats below are from mpstat. Notice the "usr" field jumps from 2 or 3 to 97
!
I also enclose results from iostat after.
Regards
Axelle
mpstat 2:
before xload:
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 132 2 4 604 220 638 21 15 11 1 2653 10 2 0 88
1 111 0 18 346 56 631 25 15 11 1 2294 8 1 0 90
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 16 0 0 610 203 658 10 10 12 0 1900 3 1 0 96
1 30 0 27 382 52 590 12 11 14 0 1676 3 1 0 96
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 528 202 480 4 10 8 0 1115 2 0 0 98
1 0 0 26 265 16 321 4 8 10 0 1196 3 1 0 96
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 499 201 428 3 5 1 0 1158 1 1 0 98
1 0 0 0 242 3 378 6 6 2 0 1270 3 0 0 96
when I launch xload:
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 160 0 18 472 168 272 14 14 16 6 617 21 1 0 78
1 198 0 29 154 36 154 34 13 20 6 774 72 2 0 26
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 610 164 574 6 4 14 0 406 1 0 0 98
1 0 0 28 286 198 63 36 4 15 0 332 99 0 0 0
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 1 613 168 436 33 19 1 0 794 20 1 0 79
1 48 0 0 254 152 422 154 20 4 0 1847 80 1 0 18
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 8 0 0 643 162 594 16 12 13 0 592 4 0 0 96
1 16 0 28 296 207 157 77 12 15 0 871 97 1 0 3
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 20 0 0 634 167 676 30 10 13 0 866 10 1 0 89
1 0 0 28 334 209 141 56 11 16 0 794 91 1 0 8
after it is loaded:
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 25 0 0 819 187 981 30 13 4 0 2211 6 2 0 93
1 40 0 1 490 167 733 60 14 16 0 2743 39 1 0 60
---------------
Now, the results from iostat (iostat -T d -xnzc 2). I've cut a few.
Note the cpu usr jumps from 2 to 50.
before I launch xload:
Thu Mar 26 21:31:50 2009
cpu
us sy wt id
3 1 0 96
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:04 2009
cpu
us sy wt id
2 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 33.0 0.0 100.5 0.0 0.0 0.3 0.6 0 2 c3d0
Thu Mar 26 21:32:06 2009
cpu
us sy wt id
5 2 0 93
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:08 2009
cpu
us sy wt id
2 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:10 2009
cpu
us sy wt id
2 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:12 2009
cpu
us sy wt id
3 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:14 2009
cpu
us sy wt id
2 1 0 96
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:16 2009
cpu
us sy wt id
3 1 0 96
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:18 2009
cpu
us sy wt id
2 1 0 97
once I run xload:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:20 2009
cpu
us sy wt id
36 1 0 63
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:22 2009
cpu
us sy wt id
50 0 0 49
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:24 2009
cpu
us sy wt id
50 1 0 49
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:26 2009
cpu
us sy wt id
50 1 0 49
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:28 2009
cpu
us sy wt id
52 1 0 47
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:30 2009
cpu
us sy wt id
34 2 0 64
after:
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:32 2009
cpu
us sy wt id
2 1 0 96
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:34 2009
cpu
us sy wt id
2 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
0.0 34.5 0.0 100.5 0.0 0.0 0.3 0.6 0 2 c3d0
Thu Mar 26 21:32:36 2009
cpu
us sy wt id
2 1 0 97
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
Thu Mar 26 21:32:38 2009
cpu
us sy wt id
3 1 0 96
extended device statistics
r/s w/s kr/s kw/s wait actv wsvc_t asvc_t %w %b device
--
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]
