thanks for your ideas… unfortunately I can be sure that’s not an browser cache
or proxy issue… and of course I simplified the problem a little bit… It’s not
me calling the empty directory tree, but an newsletter project of our company
uses shortened urls by just referencing the folder in their emails, not the
index-document to make the link look better… so the caching of 404 will appear
as soon as the first test newsletters are spread at development. Until I get my
hands on the index-document I have already plenty of calls on the webserver.
The point is, that problem disappears as soon as resin is restarted. Until then
the newsletter recipients will all see the 404 page instead of the landing page…
We are already fighting with this since a while and find no way to get around…
and for performance reasons we don’t like the idea to abandon resin as
webserver and go back to apache or any other webserver instead.
So if anyone has an idea to enlighten us what influences resins “long-term
memory” of static html-content and where to fiddle with parameters to shorten
the time to live for such content delivered by resin, please feel free to point
us into right direction.
Your idea of a customized 404 page sounds good, but it will not address the
updated hmtl-documents which won’t show their changes either until resin is
restarted… we definitely have a problem with static html-content being kind of
resistable to changes until resin is restarted.
So we still welcome ideas… or maybe someone can find same behavior?
resin-interest mailing list