Hey Brion,
Doh! It does look like the UTM code that Google Analytics uses does cookie anonymous users, will this prevent us from configuring Squid properly? The cookies won't affect the display of the pages loaded, is there a way to ignore a set of cookies that start with __utm in the context of the cache?
Travis
Message: 10 Date: Mon, 05 Dec 2005 11:32:34 -0800 From: Brion Vibber brion@pobox.com Subject: Re: [Wikitech-l] Squid configuration problem To: Wikimedia developers wikitech-l@wikimedia.org Message-ID: 439495D2.3050400@pobox.com Content-Type: text/plain; charset="iso-8859-1"
Travis Derouin wrote:
Hey,
Does anyone have any suggestions about configuring Squid in reverse proxy for MW? We're seeing that our articles aren't being cached and Squid is still making several requests per second for the same article, despite configuring Squid as instructed at meta.wikimedia.org. Here are our settings:
$wgUseSquid = true; $wgUseESI = false; $wgInternalServer = $wgServer; $wgSquidMaxage = 18000; $wgSquidServers = array('10.234.169.202'); $wgSquidServersNoPurge = array(); $wgMaxSquidPurgeTitles = 400; $wgSquidFastPurge = true;
Squid and Apache are running on different machines, and we're seeing several 200 response codes for unchanged articles from Apache, even in a small time span of a few minutes. While the apache load has been reduced, we should be seeing Squid handling more of the serving of normal article pages, instead of forwarding the request to Apache almost 100% of the time.
Can you confirm that anonymous users have no cookies? If you manually hit the server with manual HTTP can you confirm cache hits/misses?
-- brion vibber (brion @ pobox.com)
-----BEGIN PGP SIGNED MESSAGE-----
Moin,
On Tuesday 06 December 2005 19:47, Travis Derouin wrote:
Hey Brion,
Doh! It does look like the UTM code that Google Analytics uses does cookie anonymous users, will this prevent us from configuring Squid properly? The cookies won't affect the display of the pages loaded, is there a way to ignore a set of cookies that start with __utm in the context of the cache?
* Are you talking about your own mediawiki installation or wikipedia? * If wikipedia, why is Google analytics even included? * Wouldn't filtering the cookies make the inclusion of GA pointless?
I know why I personally filter GA out - including third-party JS code in a webpage is a receipe rife for desaster....
Best wishes,
Tels
- -- Signed on Tue Dec 6 20:14:11 2005 with key 0x93B84C15. Visit my photo gallery at http://bloodgate.com/photos/ PGP key on http://bloodgate.com/tels.asc or per email.
"In 1988, Jack Thompson ran against Janet Reno for DA of Dade County: Thompson's unique campaign message was that Reno was unfit for the job because, as a closeted lesbian with a drinking problem, she was great candidate for blackmail by the criminal element. Jack never explained why this remained a threat even after he exposed her 'secret'. Reno cruised at the polls."
Hey,
No, this is for our site wikiHow (http://wiki.ehow.com)
I'm trying to find out from Google whether filtering out the cookies will be pointless, I know our PHP code doesn't use them and since it's javascript code, I don't see how we would lose anything by not having GA cookies reach our Apache server.
Thanks, Travis
On 12/6/05, Tels nospam-abuse@bloodgate.com wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Moin,
On Tuesday 06 December 2005 19:47, Travis Derouin wrote:
Hey Brion,
Doh! It does look like the UTM code that Google Analytics uses does cookie anonymous users, will this prevent us from configuring Squid properly? The cookies won't affect the display of the pages loaded, is there a way to ignore a set of cookies that start with __utm in the context of the cache?
- Are you talking about your own mediawiki installation or wikipedia?
- If wikipedia, why is Google analytics even included?
- Wouldn't filtering the cookies make the inclusion of GA pointless?
I know why I personally filter GA out - including third-party JS code in a webpage is a receipe rife for desaster....
Best wishes,
Tels
Signed on Tue Dec 6 20:14:11 2005 with key 0x93B84C15. Visit my photo gallery at http://bloodgate.com/photos/ PGP key on http://bloodgate.com/tels.asc or per email.
"In 1988, Jack Thompson ran against Janet Reno for DA of Dade County: Thompson's unique campaign message was that Reno was unfit for the job because, as a closeted lesbian with a drinking problem, she was great candidate for blackmail by the criminal element. Jack never explained why this remained a threat even after he exposed her 'secret'. Reno cruised at the polls."
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux)
iQEVAwUBQ5XjnXcLPEOTuEwVAQGeEQf+K9WbLHbJ+EwbO1ou/cOl/lmEdVmYtKB1 wO9FU1g9qL4htYd+xsKdMAzNdpzlUuYinUl4f+QKYotnBiws+WV8f2X5jGUl3QxB 7UDP109U3NkPhs2rTbam3sdTnhBs4od9g+J/kifQfDDVKs5Vs7hoISc7hnqwjN5I ry0VOjlHcw9aPMHE/GaE95i04YW+rRZK/QWZCwX75ShPfuuxBjd8eOQhOXG+oEn2 7Qt04Oe1VS4q6Dz0V//qBp/W6w7vnGZ/sWw8Dx7Az3132QaJwgPNZdxStKEw0+Wc Ig4PZF7rjdOPBV9qI3UZ9qQbmhKRSu07uFg/6Jgw2+mFUzIYk30fVw== =HcpI -----END PGP SIGNATURE-----
Travis Derouin wrote:
Hey Brion,
Doh! It does look like the UTM code that Google Analytics uses does cookie anonymous users, will this prevent us from configuring Squid properly?
Probably...
The cache-control headers *should* keep private pages (logged-in versions) from getting displayed to anonymous visitors, but I'm not convinced that it can guarantee the opposite (cached anonymous pages might get shown to logged-in users) if you remove the 'Vary: Accept-Encoding, Cookie'.
The cookies won't affect the display of the pages loaded, is there a way to ignore a set of cookies that start with __utm in the context of the cache?
Sorry, I have no idea. :(
-- brion vibber (brion @ pobox.com)
wikitech-l@lists.wikimedia.org