Cache questions

[Igor Sysoev]

On Tue, Jul 14, 2009 at 11:28:15AM -0400, Jim Ohlstein wrote:

> 
> 
> Igor Sysoev wrote:
> >On Tue, Jul 14, 2009 at 09:26:22AM -0400, Jim Ohlstein wrote:
> >
> >  
> >>Igor Sysoev wrote:
> >>    
> >>>On Mon, Jul 13, 2009 at 11:09:34PM -0400, Jim Ohlstein wrote:
> >>>
> >>> 
> >>>      
> >>>>It is now logging "HIT" or "MISS" for each request in that location 
> >>>>block.
> >>>>
> >>>>It is safe to assume that with fastcgi_cache_min_uses set at 2 that for 
> >>>>every first miss shared memory is allocated, that the file is written 
> >>>>to the cache on the second miss, and that each logged "hit" is an 
> >>>>actual serve from the cache (third and subsequent requests for the 
> >>>>file)?
> >>>>   
> >>>>        
> >>>Yes. Currently, there are "MISS", "EXPIRED", "STALE", "UPDATING", and 
> >>>"HIT"
> >>>states. If you will think out a good name for state when miss is just 
> >>>counted,
> >>>but not cached, I will make this state.
> >>>
> >>>
> >>> 
> >>>      
> >>"MISS-NOT_CACHED"  fairly well tells the story, if it's not too long. 
> >>Otherwise  "MISS-NC".
> >>
> >>Or perhaps "MISS-COUNTING".
> >>    
> >
> >May be just "COUNTING" ?
> >
> >  
> So the first miss would be "MISS", subsequent misses would be 
> "COUNTING"? Or vice versa? Either way would be good for me.

No, last miss will be MISS, while the previous ones will be COUNTING.
In case "..._cache_min_uses 1" the first will be MISS, since it's
the last miss. In "..._cache_min_uses 2": COUNTING, MISS.


-- 
Igor Sysoev
http://sysoev.ru/en/