Re: [squid-users] cache object sizes

From: Amos Jeffries <squid3_at_treenet.co.nz>
Date: Fri, 03 Jun 2011 15:26:58 +1200

On 02/06/11 23:10, E.S. Rosenberg wrote:
> Hi all,
> Does having different maximum sizes for objects in memory and objects
> saved on the disk cache have a negative influence on performance or is
> using squid that way good/recommended?

Only indirectly. And there are three relevant settings. cache_dir have
one for the disks. The one you are looking at is the global limit.

Best usage and affects are determined by your 'disk' types (HDD, SSD,
ramdisk, RAM) and I/O method (UFS, AUFS, COSS) and sizes of objects your
proxy passes regularly.

Request for help:
   It has been a while since anyone presented us with benchmarking
results. ~10 years since the last full set comparing all I/O vs several
HDD disks and a range of objects.
~2 years since the last spot-check on particular disk/IO/objects-size
combo {SSD vs unspecified HDD for COSS <32KB objects}.

>
> The default squid.conf sets both to 4MB, but I'd like to also cache
> larger objects (like smaller youtube movies) so it seems to me that
> allowing larger objects in the disk cache will enable squid to cache
> the larger objects and serve them without the need for them to live in
> RAM filling the alloted RAM fast and causing more swapping then
> necessary.
>
> Currently I set the max. disk cache object size to 20MB and memory
> cache I left at 4MB.
>
> Regards and thanks,
> Eli

Squid can store YT videos just fine. Having the default big enough is a
dream shared by us upstream as well. Sadly until the YT/GV site admin
get their act together and use a friendly URL scheme we are forced to
set the limit just below their video sizes.

You can read all the gory details in
http://wiki.squid-cache.org/ConfigExamples/DynamicContent/YouTube

In fact, if you want to even try caching YT that page is a "MUST READ".

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.12
   Beta testers wanted for 3.2.0.8 and 3.1.12.2
Received on Fri Jun 03 2011 - 03:27:04 MDT

This archive was generated by hypermail 2.2.0 : Fri Jun 03 2011 - 12:00:01 MDT