RE: [squid-users] Optimizing squid

Any idea where to start looking, should I limit bandwidth per user in case some is stealing all the bandwidth, and should I change anything in my conf file.

I setup a test proxy usimg suse 8.0 + squid 2.5 stable1 and my throughput with just 4 users conneted is 150 kbps and on the old squid 13kbps (loaded)

We conect in to 2 parent squids farms in germany (Siemens)SBS
Very big and powerful (Wan pipe is 32 mbps for whole of europe, not a lot I know) and the second deals mainly with secure traffic and some http requests. These are not single parents but farms of proxys.

But our sibling here is not so fast our wan pipe to germany is 4mbps + 2x2 mbps for all our wan traffic.

Is bandwidth the problem or do we have some abusing the http proxy, sorry for the questions , as I said in inherited this, I have some knowledge of proxies but not so much on squid.

I hope I haven't too much , and thanks to all the people who answered me.

And what hit rate should I hope for.

Taken from old file
Connection information for squid:
        Number of clients accessing cache: 469
        Number of HTTP requests received: 339059
        Number of ICP messages received: 0
        Number of ICP messages sent: 0
        Number of queued ICP replies: 0
        Request failure ratio: 0.00%
        HTTP requests per minute: 262.0
        ICP messages per minute: 0.0
        Select loop called: 3424304 times, 22.671 ms avg
Cache information for squid:
        Request Hit Ratios: 5min: 6.9%, 60min: 8.1%
        Byte Hit Ratios: 5min: 15.0%, 60min: 12.8%
        Request Memory Hit Ratios: 5min: 4.6%, 60min: 4.7%
        Request Disk Hit Ratios: 5min: 50.2%, 60min: 44.4%
        Storage Swap size: 642396 KB
        Storage Mem size: 10040 KB
        Mean Object Size: 27.00 KB
        Requests given to unlinkd: 1998

Proposed config

http_port 81
ftp_user proxyadm.tma@siemens.com
cache_peer proxy.sbs.de parent 81 0 no-query default
cache_peer proxy.siemens.de parent 81 0 no-query weight=2 default cache_mem 32 MB maximum_object_size 40 MB minimum_object_size 4 KB ipcache_size 2048 fqdncache_size 2048 cache_dir ufs /usr/local/squid/cache 1024 16 256 cache_store_log none log_fqdn on dns_nameservers 163.242.244.2
refresh_pattern . 0 20% 4320
refresh_pattern -i exe$ 0 50% 999999
refresh_pattern -i zip$ 0 50% 999999
refresh_pattern -i tar\.gz$ 0 50% 999999
refresh_pattern -i tgz$ 0 50% 999999
acl all src 0.0.0.0/0.0.0.0
acl manager proto cache_object
acl localhost src 127.0.0.1/255.255.255.255
acl SSL_ports port 443 563
acl Safe_ports port 21 80 81 82 100 444 9700 8080 1997 # http
#acl Safe_ports port 21 # ftp
acl Safe_ports port 443 563 # https, snews
acl Safe_ports port 444 # snpp
acl Safe_ports port 70 # gopher
acl Safe_ports port 210 # wais
acl Safe_ports port 1025-65535 # unregistered ports
acl Safe_ports port 280 # http-mgmt
acl Safe_ports port 488 # gss-http
acl Safe_ports port 591 # filemaker
acl Safe_ports port 777 # multiling http
#acl Safe_ports port 5190 #icq
acl CONNECT method CONNECT
http_access allow CONNECT SSL_ports
http_access deny !Safe_ports
acl class1 src 163.242.237.0/255.255.255.0
acl class2 src 163.242.244.0/255.255.255.0
acl class3 src 163.242.245.0/255.255.255.0
acl class4 src 163.242.236.0/255.255.255.0
http_access allow class1
http_access allow class2
http_access allow class3
http_access allow class4
http_access allow all
acl POST method post
acl MP3 urlpath_regex \.[mM][pP][3]
http_access allow andreea MP3
http_access deny MP3 !POST
acl extension url_regex \.avi$ \.mp3$ \mp3$ .\mp3$ \.mp4$ \.mov$ \.mpeg$ \.mpg$ \.axf$ http_access allow andreea extension http_access deny extension acl network src 0/0 http_reply_access allow all miss_access allow manager miss_access allow all cache_mgr proxyadm.tma@siemens.com logfile_rotate 4 memory_pools on cachemgr_passwd pacproxy shutdown cachemgr_passwd pac info/stats/objects store_objects_per_bucket 50 client_db on buffered_logs on

Thanks thanks thanks and many many thanks Henrik

Best regards

Dave Clark
-----Original Message-----
From: Henrik Nordstrom [mailto:hno@squid-cache.org]
Sent: Donnerstag, 22. Mai 2003 12:18
To: Clark Allan Dave
Cc: squid-users@squid-cache.org
Subject: Re: [squid-users] Optimizing squid

tor 2003-05-22 klockan 09.08 skrev Clark Allan Dave:
> I have inherited a squid proxy; it is quite a loaded proxy and has
> many problems so I am trying to optimize it. At peak load it has over
> 470 users with a request per minute at about 600, and the hit rate on
> disk is about 50 % and memory is about 5% and transfers about 2.5 Gb
> of new data a day.
>
> I will upgrade the H/W this weekend with a second hard disk dedicated
> ,just to cache .This will be a 20 GB IDE disk
> Current is 8 GB drive, the disks are using UFS. Also I will increase
> ram to 768 MB.

You surely could do with a larger cache to increase the hit ratio (only
10-13% today).

As you are not having a very high request rate I am not so sure the
bottleneck is your hardware.. but upgrading the hardware never hurts
performance.

I'd recommend you to spend some time to study the cause of the problems
you are having and not betting everything on the upgrade.

> Also is it worth upgrading to squid 2.5 stable, as we run 2.5 Devel.

DEVEL releases SHOULD NEVER be used in production. These are ment for
testing and development only.

Regards
Henrik

-- 
Donations welcome if you consider my Free Squid support helpful.
https://www.paypal.com/xclick/business=hno%40squid-cache.org
Please consult the Squid FAQ and other available documentation before
asking Squid questions, and use the squid-users mailing-list when no
answer can be found. Private support questions is only answered
for a fee or as part of a commercial Squid support contract.
If you need commercial Squid support or cost effective Squid and
firewall appliances please refer to MARA Systems AB, Sweden
http://www.marasystems.com/, info@marasystems.com