Re: Question about efficient URL prefix search

[Leif Hedstrom <zw...@apache.org>]

On 1/29/13 8:57 AM, James Peach wrote:
> On 29/01/2013, at 1:11 AM, Bowei Du <bd...@instartlogic.com> wrote:
>
>> Hi all,
>>
>> I am looking for a way to quickly get a list of all objects stored in the
>> ATS cache that start with a given string prefix (for example, given "
>> foo.com/bar", returns objects {foo.com/bar123, foo.com/bar456,  ...}). As
>> far as I can tell, the actual URL of the cached object is not stored in
>> memory and this operations requires scanning the volumes and reading the
>> metadata for each HTTP object from disk.
> I think that you'd have to build an index in memory or in the cache. I don't think you can iterate the HTTP cache from a plugin, so you might have to hack the core :(
>


It should be possible to traverse the entire cache, but it'll be incredibly 
expensive and slow on large caches. See TSCacheScan() and the cache_scan 
plugin for more details on scanning the cache.

-- Leif