How to handle the "page just expired in front-end cache" case?

[Bertrand Delacretaz <bd...@apache.org>]

I'd like to pick the community's brains on this use-case:

1. Cocoon-based pages are cached by mod_cache

2. A request for page P comes, page P is invalid in cache, request goes 
to Cocoon

3. During the time it takes to regenerate page P, more requests arrive 
for page P

What happens at this point?

My guess (I'm still investigating and testing) is that the additional 
requests also go through the cache, causing (too) many concurrent 
requests for page P on the Cocoon instance.

The right thing to do would be for the front-end cache to know that 
page P is being regenerated, and have the additional requests wait 
until it's updated in the cache.

Does anyone know if this is how mod_cache (in apache2) is supposed to 
work?

As I said, I'm still investigating, I shall have more evidence soon, 
but if someone already knows how this works, I'm all ears!

-Bertrand