Re: [squid-users] very slow browsing and page is not displaed properly

goody goody wrote:
> Hi,
>
> Squid GURUs, Your response is required, please.
>
> Regards,
> .Goody.
>
>
> ----- Original Message ----
> From: goody goody <thinkodd_at_yahoo.com>
> To: squid-users_at_squid-cache.org
> Sent: Fri, May 21, 2010 1:52:23 AM
> Subject: Re: [squid-users] very slow browsing and page is not displaed properly
>
> Dear Members,
>
> In addition to below information, I have added some more info regarding machine hardware and platform.
>
> RAM = 4 GB
> Processors = 4
> HDDs SATA having implemented RAID-5
>
> Running on VMWARE ESXi 3.5.
>
> Should you need any info, pls let me know.
>
> Waiting for your expert opinion, please.

http://wiki.squid-cache.org/SquidFaq/RAID

45GB cache on RAID-5. ouch.

If you are really absolutely forced to use RAID at all, go for RAID-1.

For best performance go with JBOD and do away with half the physical
level IO.

I'd also recommend an OS other than *BSD for AUFS. There are some write
problems on BSD that apparently slow it down.

>
> ----- Original Message ----
> From: goody goody <thinkodd_at_yahoo.com>
> To: squid-users_at_squid-cache.org
> Sent: Thu, May 20, 2010 4:31:21 PM
> Subject: [squid-users] very slow browsing and page is not displaed properly
>
> Hi,
>
> Version information and some statistics collected by me are as below. At times, my users complain the browsing becomes deadly slow and we page like yahoo, after much delay is displayed scattered and pictures are not visible rather "X" sign is displayed and after few times refresh screen becomes better.
>
> proxy-br# uname -a
> FreeBSD proxy-br 0-RELEASE FreeBSD 8.0-RELEASE #0: Sat Nov 21 15:48:17 UTC 2009 root_at_almeida.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC i386
>
> --------------------------------------------------------------------------------------------------------------------
>
> proxy-br# /usr/local/squid27/sbin/squid -v
> Squid Cache: Version 2.7.STABLE9
> configure options: '--prefix=/usr/local/squid27' '--enable-async-io' '-enable-storeio=aufs,coss'
> '--enable-removal-policies=heap,lru' '--enable-snmp' '--with-openssl=/opt/ssl' '--enable-wccp'
>
> --------------------------------------------------------------------------------------------------------------------
>
> proxy-br# iostat -c 5 -w 3
> tty da0 pass0 cpu
> tin tout KB/t tps MB/s KB/t tps MB/s us ni sy in id
> 0 138 13.88 2 0.03 0.00 0 0.00 4 0 1 0 95
> 0 140 11.00 1 0.01 0.00 0 0.00 11 0 5 1 83
> 0 133 11.00 1 0.01 0.00 0 0.00 16 0 5 1 78
> 0 86 16.00 0 0.01 0.00 0 0.00 13 0 4 1 82
> 0 132 3.07 5 0.01 0.00 0 0.00 14 0 4 1 80
>
> --------------------------------------------------------------------------------------------------------------------
> proxy-br# vmstat
> procs memory page disks faults cpu
> r b w avm fre flt re pi po fr sr da0 pa0 in sy cs us sy id
> 1 0 0 924M 154M 20 0 0 0 6 1 0 0 189 1178 1366 4 1 95
>
> --------------------------------------------------------------------------------------------------------------------
> proxy-br# systat
>
> /0 /1 /2 /3 /4 /5 /6 /7 /8 /9 /10
> Load Average ||
>
> /0% /10 /20 /30 /40 /50 /60 /70 /80 /90 /100
> root idle XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
> root idle XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
> squid squid XXXXXXXXXXXXXXXXX
> root kernel X
>
>
> my squid.conf is as below
> --------------------------------------------------------------------------------------------------------------------
> http_port 3128
>
> hierarchy_stoplist cgi-bin ?
> acl QUERY urlpath_regex cgi-bin \?
> no_cache deny QUERY

"no_cache" is dead. Drop the "no_" part of it.

Also the QUERY acl is now deprecated. You can drop those above lines
entirely and add a refresh_pattern instead:

  refresh_pattern -i (/cgi-bin/|\?) 0 0% 0

(needs to go just about the "refresh_pattern . " line).

> cache_mem 256 MB
> visible_hostname pxy
> #negative_ttl 0
>
> acl PURGE method PURGE
> acl localhost src 127.0.0.1
> http_access allow PURGE localhost
> http_access deny PURGE
>
>
> cache_replacement_policy heap LFUDA
> memory_replacement_policy heap GDSF
>
> cache_dir aufs /cache 45000 16 256
>
> cache_store_log /dev/null #/var/log/squid27/store.log

Drop the line above. It's both broken syntax (there are no comments
allowed at the end), and unused as the next line below erases it.

> cache_store_log none
> cache_swap_low 80
> cache_swap_high 90
> cache_log /var/log/squid27/cache.log
> cache_access_log /var/log/squid27/access.log
>
> half_closed_clients off
>
>
> ...
> ...acl...
> ....
> .....
>
> #always_direct allow myiplist
> cache_mgr xxxx_at_xxxxxxxx
> cache_effective_user squid
> cache_effective_group squid
> logfile_rotate 0
> buffered_logs on
> nonhierarchical_direct off
> prefer_direct off
> ie_refresh on
> ftp_list_width 32
> ftp_passive on
> ftp_sanitycheck on
> ftp_telnet_protocol on
>
>
> emulate_httpd_log on

Use the "common" log format on each access_log line instead of this
deprecated option.

ie. cache_access_log /var/log/squid27/access.log common

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.3