Marcin Soltysiak schrieb:
Kernel: 2.6.18-4-686
Try upgrading to 2.6.20. I've read about some issues with file handlers with
2.6.18. After upgradiong I reduced iowait% significantly.
Apache: 2.2.3-4
MySQL: 5.0.32-Debian_7etch1-log
And did you make any diagnosis about mysql tuning? Like threads cache, queries
cache slow queries etc? I gained a lot by that.
mpstat:
12:01:15 CPU %user %nice %sys %iowait %irq %soft %steal
%idle intr/s
12:01:15 all 2,43 0,02 0,51 9,98 0,12 0,01 0,00
86,93 283,67
uptime:
12:01:18 up 25 days, 2:22, 1 user, load average: 2.67, 5.46, 6.00
And how comes that you got high load while 85% idling without any iowaits? Try
observig this with top. It will show both load and CPU's usage same time.
Here' the output from top:
28709 www-data 16 0 143m 72m 16m D 0.7 14.3 2:51.47 apache2
31042 www-data 16 0 109m 46m 16m D 0.7 9.2 1:29.60 apache2
113 root 10 -5 0 0 0 D 0.3 0.0 18:04.26 kswapd0
28688 www-data 16 0 217m 43m 16m D 0.3 8.7 3:59.65 apache2
28707 www-data 18 0 123m 59m 16m D 0.3 11.7 2:10.25 apache2
32242 root 15 0 2228 1128 860 R 0.3 0.2 0:00.04 top
1 root 15 0 1944 84 64 S 0.0 0.0 0:16.39 init
2 root RT 0 0 0 0 S 0.0 0.0 0:00.00
migration/0
3 root 34 19 0 0 0 S 0.0 0.0 0:00.07
ksoftirqd/0
4 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 events/0
5 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 khelper
6 root 10 -5 0 0 0 S 0.0 0.0 0:00.26 kthread
9 root 10 -5 0 0 0 S 0.0 0.0 0:09.17 kblockd/0
10 root 20 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid
77 root 10 -5 0 0 0 S 0.0 0.0 0:00.00 kseriod
114 root 20 -5 0 0 0 S 0.0 0.0 0:00.00 aio/0
815 root 10 -5 0 0 0 D 0.0 0.0 0:58.63 kjournald
the corresponding uptime readings are:
10:51:22 up 26 days, 1:12, 1 user, load average: 8.27, 6.79, 5.24
I could try updatig the kernel and tuning mysql, but the point is:
Before updating Midgard, the system ran smoothly, now the response times
are between 10 and 20 times higher. Even if tuning can make the system
400% faster, it would still be way slower than it was with 1.8.2.2.
Another interesting thing is that when I restart Apache, load times are
back to normal. But after a couple of hours, they return to the insanely
high values metioned before, but I have no idea what triggers it.
Bye,
Andreas
Solt
------------------------------------------------------------------------
_______________________________________________
user mailing list
[email protected]
http://lists.midgard-project.org/mailman/listinfo/user
begin:vcard
fn:Andreas Flack
n:Flack;Andreas
org:CONTENT CONTROL Berlin GbR
adr;dom:;;Ebertystr. 20;Berlin;Berlin;10249
email;internet:[EMAIL PROTECTED]
tel;work:+49 (0) 177 565 19 13
tel;fax:+49 (0) 30 639 018 64
url:http://www.contentcontrol-berlin.de
version:2.1
end:vcard
_______________________________________________
user mailing list
[email protected]
http://lists.midgard-project.org/mailman/listinfo/user
