Re: [squid-users] High CPU usage problem on Squid 2.6 STABLE9

From: Robert <vipit@dont-contact.us>
Date: Tue, 30 Jan 2007 20:51:49 +0100

Andrew Miehs napisał(a):
> Hi Robert,
>
> This box seems to be running something a little MORE than squid -
> could this be a firewall perhaps?
Yes it is iptables firewall,
>
> The huge number of interrupts then seem ok for this sort of a setup -
> How many Megabits/ second are you pushing through this box?!
>
My connection speed is 45 Mbit max, but real traffic is about 30-35 Mbit
including P2P.
>
> Sorry to be a pain - I have more quesions...
>
> Is this box really an Intel box, or does it have Opterons installed?
> (hostname name = sunfire)
This is Sun SunFire V65x server
http://www.sun.com/servers/entry/v65x/
>
> what does
> mpstat -P ALL 5
> output ?
>
Output with transparent squid traffic enabled:
20:33:05 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:10 all 1,55 0,00 1,00 12,00 1,35 21,95
62,15 8930,94
20:33:10 0 3,79 0,00 1,20 0,60 5,39 87,82
1,20 8930,94
20:33:10 1 2,00 0,00 1,60 10,98 0,00 0,00
85,43 0,00
20:33:10 2 0,00 0,00 0,60 29,54 0,00 0,00
69,86 0,00
20:33:10 3 0,40 0,00 0,80 6,79 0,00 0,00
91,82 0,00

20:33:10 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:15 all 1,00 0,00 1,00 13,20 1,30 22,95
60,55 9004,01
20:33:15 0 2,20 0,00 1,20 1,00 5,21 89,98
0,40 9004,41
20:33:15 1 1,20 0,00 2,00 34,87 0,00 2,00
60,12 0,00
20:33:15 2 0,20 0,00 0,00 6,81 0,00 0,00
92,99 0,00
20:33:15 3 0,40 0,00 0,60 10,22 0,00 0,00
88,98 0,00

20:33:15 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:20 all 1,70 0,00 1,00 5,90 1,80 22,00
67,60 8946,40
20:33:20 0 4,40 0,00 0,60 0,20 7,20 87,20
0,40 8946,20
20:33:20 1 1,40 0,00 1,80 12,40 0,00 0,40
84,00 0,00
20:33:20 2 0,20 0,00 1,40 5,60 0,00 0,20
92,60 0,00
20:33:20 3 0,60 0,00 0,40 5,60 0,00 0,00
93,60 0,00

20:33:20 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:25 all 1,60 0,00 1,05 5,80 1,50 22,00
68,05 8913,60
20:33:25 0 3,80 0,00 1,40 0,00 6,00 87,60
1,20 8913,60
20:33:25 1 2,40 0,00 1,80 9,40 0,00 0,40
86,20 0,00
20:33:25 2 0,20 0,00 0,40 6,80 0,00 0,00
92,60 0,00
20:33:25 3 0,20 0,00 0,60 6,80 0,00 0,20
92,20 0,00

20:33:25 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:30 all 3,45 0,00 0,85 3,30 1,60 19,90
70,90 9180,44
20:33:30 0 12,57 0,00 1,20 0,00 6,39 79,24
0,60 9180,44
20:33:30 1 0,80 0,00 1,20 1,20 0,00 0,00
96,41 0,00
20:33:30 2 0,20 0,00 0,40 0,80 0,00 0,00
98,40 0,00
20:33:30 3 0,20 0,00 0,60 11,38 0,00 0,20
87,43 0,00

20:33:30 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:35 all 4,00 0,00 1,95 8,40 1,65 19,05
64,95 8526,40
20:33:35 0 11,60 0,00 3,20 0,20 6,60 75,80
2,60 8526,20
20:33:35 1 4,00 0,00 2,80 10,60 0,00 0,40
82,40 0,00
20:33:35 2 0,00 0,00 0,80 7,20 0,00 0,00
92,00 0,00
20:33:35 3 0,20 0,00 1,00 15,80 0,00 0,00
83,00 0,00

20:33:35 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:40 all 3,35 0,00 2,10 10,24 1,35 18,84
64,12 8767,80
20:33:40 0 12,40 0,00 3,60 0,20 5,40 75,20
3,20 8768,20
20:33:40 1 0,80 0,00 4,00 35,20 0,00 0,00
59,80 0,00
20:33:40 2 0,20 0,00 0,20 2,00 0,00 0,00
97,40 0,00
20:33:40 3 0,00 0,00 0,60 3,20 0,00 0,00
96,20 0,00

20:33:40 CPU %user %nice %sys %iowait %irq %soft
%idle intr/s
20:33:45 all 3,90 0,00 1,55 10,15 1,90 19,60
62,90 8399,60
20:33:45 0 10,40 0,00 2,20 0,80 7,60 76,00
3,00 8399,20
20:33:45 1 4,40 0,00 2,20 16,60 0,00 2,40
74,40 0,00
20:33:45 2 0,40 0,00 0,60 5,20 0,00 0,00
93,80 0,00
20:33:45 3 0,40 0,00 1,00 18,00 0,00 0,00
80,40 0,00

> Is the latency on sites in the cache? or 'new' sites?
Hmm i'll see these but eg. openning googll site with squid enabled take
1-3 sec.with squid disabled take max 1 sec
> Could this be a DNS issue? Are you running a pair of caching DNS
> servers on site?
> or are they based at an ISP?
>
I have installed DNScache local server, all dns works ok.
> Are you running as a 'transperant' proxy? or have the users entered
> the proxy details per
> hand or per proxy.pak.... ?
>
transparent with iptabes redirect command from port 80 to port 3128.

> I would probably run aufs rather than diskd - but this should not make
> a real difference..
> You have also only configured 3Gs of cache (3x 1G)...
> You may want to try disabling the disk cache altogether and see how
> that works - you need to
> compile with -enable-storeio=diskd,aufs,null
>
cache_dir null ? in squid.conf ?
> I would also increase my cache size to perhaps, 2G.. (How much RAM
> does the machine have?)
>
2 GB RAM
>
> Sorry that I am not of more help,
>
Thanks a lot
> Cheers
>
> Andrew
>
Regards
Robert
Received on Tue Jan 30 2007 - 12:52:03 MST

This archive was generated by hypermail pre-2.1.9 : Thu Feb 01 2007 - 12:00:01 MST