|
Re: Does key length require storage?: msg#00074web.cache.memcached
On Mon, Jan 24, 2005 at 04:08:52PM -0800, Kevin A. Burton wrote: > Oh ... wow... the class size is dependent on both the key AND the data? In the current storage model, yes, but that's not guaranteed to remain unchanged and should not be relied upon. In the current storage model, the server takes a (-n almost) fixed-size header, concatenates it with the key and the data, and uses the resulting chunk of memory as one item to be size-classed and stored. > If so then does it really make much sense to have a 2^3 size class? No, there are never any actual items in that class in practice, and same goes for 2^4; however, all they take up is just one tiny preallocated struct per each, so it's not really a problem. -- avva "There's nothing simply good, nor ill alone" -- John Donne |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Re: Does key length require storage?: 00074, Kevin A. Burton |
|---|---|
| Next by Date: | apr_memcache 0.6.1 released: 00074, Paul Querna |
| Previous by Thread: | Re: Does key length require storage?i: 00074, Kevin A. Burton |
| Next by Thread: | apr_memcache 0.6.1 released: 00074, Paul Querna |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |