At the risk of answering my own question and for future reference. Found out the problem wasn't the ram cache after all. We had been using different style sheets depending on the browser you view each page with. Unfortunately, I had forgotten this fact. The header checks the string and then embeds the call to the appropriate style sheet in the page as it's generated. Of course, whomever got there first loaded the cache, and if you viewed the page from another browser, it looked like it was "corrupted". Settled on going back to a single Style sheet which took care of the problem. Also tested caching differing versions of pages dependent on the browser. Worked great, but of course that can eat up more memory per page. Thanks, -Chris ------------------------------ Chris Kratz chris.kratz@vistashare.com ----- Original Message ----- From: "Chris Kratz" <chris.kratz@vistashare.com> To: <zope@zope.org> Sent: Thursday, November 29, 2001 9:04 AM Subject: Question about Ram Cache
Hello All,
Recently, we have been doing a lot more caching of various parts of our website. We have noticed two times over the last week that after a restart and a long period of relative inactivity (overnight) that one of the pages has gotten corrupted. Invalidating the page in the ram cache and forcing it to reload takes care of the problem. The first time the HTML was corrupted (parts were missing) and in the second case, none of the images (which are also cached) would show even though those same images on other areas of the website displayed correctly.
The page is quite simple and includes only static text with a header and footer that is generated with a date and the generation time for the page. The only REQUEST variable is AUTHENTICATED_USER which means that all anonymous views get the same cached page. Here are the other settings we are using:
Threshold entries: 1000 Max Age: 3600 Cleanup interval: 300
Also, there are no Names from the DTML namespace currently being used as cache keys on the page in question.
Since this is such a basic application of the Ram Cache, I am a little surprised that we would be seeing this type of behavior. Anyone have any ideas why this might be happening?
RH 7.2, Zope 2.4.1, Python 2.1.1
Thanks,
-Chris ------------------------------ Chris Kratz chris.kratz@vistashare.com