Avinash Rao wrote:
On Thu, Sep 10, 2009 at 10:54 AM, Avinash Rao <avinash.aol@xxxxxxxxx> wrote:
On Wed, Sep 9, 2009 at 5:45 PM, Amos Jeffries <squid3@xxxxxxxxxxxxx> wrote:
Avinash Rao wrote:
Dear all,
I am sure this question would have been posted many times. I read a
few threads, but my requirement or setup is a bit different. I am
running Squid2.6stable18 on Ubuntu Server 8.04 Server 64-bit installed
on Sun Fire X4150 Server with 8GB RAM + 8 SAS HDD's - RAID 5 + 2 Quad
Core Intel Xeon Processors.
I have both LTSP and WinXP clients using Squid. Many times i have
noticed, squid stops responding, the browser keeps trying to connect
and i don't see any error in cache.log or store.log. My only way out
is to restart squid.
Usually means Squid is doing something with its on-disk storage.
I read through cache_mem, my current configuration is 100MB, i
changed it to 128MB, but squid couldn't start saying the cache_mem is
more than cache_disk size.
Why does this happen and what is the recommended configuration for the
hardware i have?
You are missing a cache_dir option. Which means Squid-2 will be using the
default and very inefficient 100 MB ufs formatted storage.
Squid-3.1 and later are the only releases of Squid where cache_dir can be
fully removed from the config.
Either define a cache_dir to use the disks or setup the "null" cache_dir
type for memory-only storage.
squid.conf: Please donot compare this config to my other posts, as
this is on a different server and it has a very basic configuration.
visible_hostname sunserver
hierarchy_stoplist cgi-bin ?
acl QUERY urlpath_regex cgi-bin \?
no_cache deny QUERY
Drop 'no_cache' use instead:
cache deny QUERY
hosts_file /etc/hosts
http_port 10.10.10.10:3128
refresh_pattern ^ftp: 1440 20% 10080
refresh_pattern ^gopher: 1440 0% 1440
refresh_pattern . 0 20% 4320
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 to_localhost dst 127.0.0.0/8
acl SSL_ports port 443 563
acl Safe_ports port 80 # http
acl Safe_ports port 21 # ftp
acl Safe_ports port 443 563 # https, snews
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 631 # cups
acl Safe_ports port 777 # multiling http
acl Safe_ports port 901 # SWAT
acl Safe_ports port 993 # IMAP
acl Safe_ports port 587 # SMTP
acl Safe_ports port 22 # SSH
acl purge method PURGE
acl special_urls url_regex "/etc/squid/squid-noblock.acl"
acl extndeny url_regex -i "/etc/squid/blocks.files.acl"
acl malware_block_list url_regex -i "/etc/squid/malware_block_list.txt"
acl badurl url_regex -i teen orkut youtube sex mp3 mp4
acl lan src 10.10.10.0/24
acl stud ident_regex babu
acl download method GET
acl CONNECT method CONNECT
cache_mem 100 MB
ident_lookup_access allow all
deny_info http://malware.hiperlinks.com.br/denied.shtml malware_block_list
http_access allow manager localhost
http_access deny manager
http_access allow purge localhost
http_access allow special_urls
http_access deny extndeny download
http_access deny extndeny
http_access deny purge
http_access deny !Safe_ports
http_access deny CONNECT !SSL_ports
http_access deny badurl
http_access deny malware_block_list
http_access allow localhost
http_access allow lan
http_access deny all
http_reply_access allow all
icp_access allow all
coredump_dir /var/spool/squid
Many Thanks
Avinash
--
Please be using
Current Stable Squid 2.7.STABLE6 or 3.0.STABLE18
Current Beta Squid 3.1.0.13
Thank you all for the information. I read the documentation.
To begin with I have set cache_dir to 256MB and cache_mem also to 256
MB. Which takes the precedence? I guess the pages are stored/retrieved
from the RAM. Does this setting mean that, once 256MB is reached in
RAM it will start storing the pages in cache_dir?
Thanks,
Avinash
Hi Guys,
I just tested it, but it was pretty slow, the page was not loading
only. Strange coz,
the cache_mem is increased to 256MB and its still slow? How does this work?
Everything gets stuffed into cache_mem first.
- objects which are labeled as too large for memory, or those
unlabeled which grow too large, get pushed to disk. This happens even if
not completely downloaded yet, which can be one cause of slowness on big
stuff.
- objects which have not been used longest get pushed to disk when
more memory space is needed.
- all memory objects are pushed to disk during shutdown, restart,
reconfigure.
According to the squid Documentation, i recommended size for cache_dir
is 7000 or 8000!
I set cache_dir to null /tmp and i could see a lot of difference and
it worked pretty well.
Seems to me that the issue was some things being sent to disk storage.
Combined with severe speed problems reading/waiting from that disk storage.
You said you have RAID5 so I'm not surprised its slow.
RAID and Squid do NOT play well together. Squid treats the disk as
random-access media using as many access threads as it can obtain in
parallel. Doing 4KB chunk operations.
Amos
--
Please be using
Current Stable Squid 2.7.STABLE6 or 3.0.STABLE18
Current Beta Squid 3.1.0.13
