Re: [squid-users] Caching YouTube content

On 02/09/11 06:03, Kevin Wilcox wrote:
> On Thu, Sep 1, 2011 at 13:02, Benjamin wrote:
>
>> Anyone used store_url from squid version 2.7?
>>
>> Any feedback or suggestion of this article?
>>
>> If someone tried it, please share your suggestions.
>
> I'm testing it but it's a long-term test phase as I can only test for
> a few months out of the year on any sizable population.
>
> On a small test deployment I can get the content to cache - and
> retrieve - using store_url and a perl handler, at least for youtube
> content, but the regex needs some work as it's *highly* inefficient,
> not at all ready to scale out beyond 100 or so users (which is about
> 5% of the population I need to handle on a single proxy) and requires
> the URL to be *exactly* alike between viewings (you know all those
> little&related= variations? they cause multiple downloads).

What do you mean? The regex we have is supposed to erase all the
volatile parameter pieces. So storeurl produces some fake domain with
the video unique ID and HD resolution type as path.

Could you please check what you have/know against our details and add a
comment of your own to the discussion page?
http://wiki.squid-cache.org/ConfigExamples/DynamicContent/YouTube

>
> If I get it squared away I'll get some documentation and examples together.

YT change things every few months, we have found waiting for a full
result is not worth it. That wiki "tutorial" is being used as a living
document with collaborative research into the remaining problems and
solutions. You are welcome to join in.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.15
   Beta testers wanted for 3.2.0.10