Re: Race in svn_atomic_namespace__create

On Tue, Oct 30, 2012 at 7:18 PM, Stefan Sperling <stsp_at_elego.de> wrote:

> On Tue, Oct 30, 2012 at 05:58:20PM +0000, Philip Martin wrote:
> > Another approach would be to create the shared memory created from some
> > other, long-lived, process. The user would have to run this process to
> > enable caching. To handle a large number of repositories this would
> > probably have to be some sort of daemon.
>
> Can we require an auxilliary process such as memcached to hold
> cached revprop data instead of caching it in-process?
> Or would that perform worse or equal to no caching at all?
>

Latency would probably be a major issue here even
if we used some RPC or socket interface. In that case,
we might as well read the file content from the OS'es
file cache. It is at least unclear whether we could still
be much faster than that.

-- Stefan^2.

-- 
Certified & Supported Apache Subversion Downloads:
*
http://www.wandisco.com/subversion/download
*