Hi Lawrence I think that's the most extensive list of refresh patterns I've seen in one place, for a forward proxy. Props. Is anyone else using a collection like this and care to comment on its performance / viability? Dan On 29 Apr 2014, at 2:05 pm, Lawrence Pingree <geekguy@xxxxxxxxxxxx> wrote: > Try using my refresh patterns: > http://www.lawrencepingree.com/2014/01/01/optimal-squid-config-conf-for-3-3- > 9/ > > > > > Best regards, > The Geek Guy > > Lawrence Pingree > http://www.lawrencepingree.com/resume/ > > Author of "The Manager's Guide to Becoming Great" > http://www.Management-Book.com > > > > -----Original Message----- > From: Amos Jeffries [mailto:squid3@xxxxxxxxxxxxx] > Sent: Monday, April 28, 2014 10:15 AM > To: squid-users@xxxxxxxxxxxxxxx > Subject: Re: how to use refresh_pattern correct > > On 29/04/2014 2:02 a.m., tile1893 wrote: >> Hi, >> >> i'm running squid on openwrt and i want squid to cache all requests > which >> are made. >> I think that this is done by defining refresh_pattern in squid config. >> But in my opinion no matter how i config them, they are always be > ignored by >> squid and will never be used. >> >> for example: >> refresh_pattern www 5000 100% 10000 override-expire > override-lastmod >> ignore-reload ignore-no-store ignore-must-revalidate ignore-private >> ignore-auth store-stale >> >> or: >> refresh_pattern www 1200 100% 6000 override-expire >> >> But they both dont work. >> Any idea how to configure squid that it is caching every request?! Do >> I > have >> to enable those refresh_patterns somehow?! > > FYI: Caching everything is not possible. HTTP protocol requires at least > some non-cached traffic just to operate. > > Now that your expectations have been lowered ... > > *correct* usage is not to have any of the override-* or ignore-* options at > all. But correct and practical are not always the same. Use the options if > you are required to, but only then. > > There is also a very large diference betwen HTTP/1.0 caching and > HTTP/1.1 caching you need to be aware of. In HTTP/1.0 there was HIT/MISS and > very little else. In HTTP/1.1 there is also revalidation (304, REFRESH, IMS, > INM) which is caching the [large] bodies of objects while still sending the > [small] headers back and forth - giving the best of both worlds. > > > So tile1893... > what version of Squid do you have? > how are you testing it? > what makes you think its not caching? > how much cache space do you have? > what are your maximum object limits? > what order is your cache, store and object related config options? > what traffic rate (requests per second/minute) are you serving? > what does redbot.org say about the URLs you are trying to cache? > > (Maybe more later but that should do for starters.) > > Amos > > >
