-----Original Message----- From: Terry <td3...@gmail.com> Sent: Thursday, August 13, 2009 9:28 PM To: Amos Jeffries <squ...@treenet.co.nz> Cc: squid-users@squid-cache.org <squid-users@squid-cache.org> Subject: Re: [squid-users] caching dynamic image content
On Thu, Aug 13, 2009 at 6:36 PM, Amos Jeffries<squ...@treenet.co.nz> wrote: > Terry wrote: >> >> On Thu, Aug 13, 2009 at 9:32 AM, Terry<td3...@gmail.com> wrote: >>> >>> On Thu, Aug 13, 2009 at 8:10 AM, Amos Jeffries<squ...@treenet.co.nz> >>> wrote: >>>> >>>> Terry wrote: >>>>> >>>>> Hello, >>>>> >>>>> I don't have squid implemented yet. >>>>> >>>>> I am researching a web architecture issue I am seeing with a site. >>>>> Squid may be a bandaid for what I think may be some poor development >>>>> architecture decisions. There are concerns that the site is written >>>>> in a way that browsers and reverse proxies cannot cache it >>>>> appropriately. And these aren't my concerns by the way. We also have >>>>> A10 load balancers in house that do some caching. They said they >>>>> can't cache this content. I don't want to go into their reasoning >>>>> because I don't believe it. >>>>> >>>>> Here's an example of an image as seen from the client. I pulled this >>>>> right out of my firefox memory cache: >>>>> http://foo.domain.com/Image.aspx?i=db1edbcd-2375-4bae-b33f-a53ced60deed >>>>> >>>>> 1. If it's in the memory cache, can I assume that browsers and proxies >>>>> can cache it? Also, I never saw these objects in my disk cache. Not >>>>> sure if that's significant or not. >>>> >>>> No. The browser has additional information such as who is logged in and >>>> whether your session with the website is the same. They are also allowed >>>> to >>>> cache objects personal to you. >>>> >>>> Proxies and caches only have the URL and some other limited data to base >>>> the >>>> checking on. If there is any chance it was a private object it will not >>>> be >>>> cached naturally. >>>> >>>>> 2. Does firefox still interpret this as an image and cache it as one >>>>> or is this considered dynamic content that may be problematic? >>>> >>>> Not enough information to even guess. What headers are present? Does the >>>> website require login? does the same image ever change URL (including >>>> the >>>> query string) and why/when/how often? are alternative image formats >>>> available at the exact same URL? >>>> >>>> Any one of those answers may make the object non-cacheable by shared >>>> proxies. >>>> >>>>> I think that's enough information to start a conversation. Thanks for >>>>> any insight! >>>> >>>> foo.domain.com does not resolve here so I can't verify the object. >>>> Please pick some of the URLs and enter them into http://www.redbot.org >>>> for >>>> review of cacheability. >>>> >>>> Amos >>>> -- >>>> Please be using >>>> Current Stable Squid 2.7.STABLE6 or 3.0.STABLE18 >>>> Current Beta Squid 3.1.0.13 >>>> >>> >>> Thank you both for replying. I haven't messed with squid and caching >>> for 5+ years and its all slightly coming back to me. The identifier >>> in the URL is not unique based upon the session of the user. >>> https://foo.domain.com/Image.aspx?i=db1edbcd-2375-4bae-b33f-a53ced60deed >>> >>> the i=db1edbcd-2375-4bae-b33f-a53ced60deed is a unique identifier for >>> the image and its size. Based on that, it should be cacheable but >>> the developers are setting it to nocache for some reason. I am >>> guessing they reused some code for other dynamic content and failed to >>> see this aspect. >>> >> >> Just to further prove its not being cached, here's the header: >> >> >> GET /Pic.aspx?i=db1edbcd-2375-4bae-b33f-a53ced60deed HTTP/1.1 >> Host: foo.domain.com >> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; >> rv:1.9.1.2) Gecko/20090729 Firefox/3.5.2 (.NET CLR 3.5.30729) >> Accept: image/png,image/*;q=0.8,*/*;q=0.5 >> Accept-Language: en-us,en;q=0.5 >> Accept-Encoding: gzip,deflate >> Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7 >> Keep-Alive: 300 >> Connection: keep-alive >> Referer: https://foo.domain.com/ >> Cookie: com.domain.foo_SessionID=kua4ew454kjodsjpdlojdu55 >> Cache-Control: max-age=0 > > The reply headers are kind of more important for the reply is what gets > stored. > > > Um, just in case it has not already been done... > check that the QUERY ACL and its "cache deny" rule are removed from your > squid.conf and "refresh_pattern -i (/cgi-bin/|\?) 0 0% 0" is added on the > line above the dot '.' refresh pattern. > > > Amos > -- > Please be using > Current Stable Squid 2.7.STABLE6 or 3.0.STABLE18 > Current Beta Squid 3.1.0.13 > Thank you. Yes, that was done. I found my resolution through the assistance of squid. In short, devs were specifying no-cache, which I figured out quickly, but without making those changes, I wanted to see the benefits of caching. I saw a 350% performance increase by throwing squid in front in reverse mode. Thanks all for your replies.
