Using memcached to cache subversion meta-data

[Pavel Francirek]

I tried to write little path to read output in blocks (4096 bytes) but
there was not significant performance improvement.

We are using memcache as memory storage for 5500+ online users web chat
written in Python (from Monday :-))

We are now testing encapsulation of some C libraries into Python.

Pavel


On Thu, 2005-02-03 at 04:40 -0800, Brad Fitzpatrick wrote:
> It's great to have a Python developer using Memcached... your next project
> could be fixing up the Memcached python client.  :) Or at least
> sanity-checking the code for us.
> 
> If I recall, the performance of it is pretty horrid... it does tons of
> 1-byte reads when reading responses from the server?  Evan was always
> concerned about this:
> 
>     def readline(self):
>         newlines = 0
>         buf = ''
>         while newlines < 2:
>             char = self.socket.recv(1) # XXX does this buffer or is this slow?
>     ....
> 
> Wanna look it over?
> 
> - Brad
> 
> 
> On Thu, 3 Feb 2005 christopher at baus.net wrote:
> 
> > Hi all,
> >
> > I just thought I would let you know that I've found a unique use for
> > memcached.  I'm using it to cache subversion meta-data queries.  This
> > isn't a large scale scalability problem like those running massive sites
> > like livejournal and meetup.  It simply solves a problem where a query
> > doesn't change often, but takes a long time to complete.
> >
> > I have to say it is one of the easiest Open Source projects I've ever
> > used.  It literally took about a 1/2 an hour to add support.  A custom
> > solution would probably have taken days.  More information on that project
> > is here:
> >
> > http://baus.net/svnpyblosxom
> >
> > I'm so happy with it, that I'm considering using it in another context all
> > together.  I working on a project to cache HTTP proxy rules for fast
> > retrieval.  I planned on using shared memory, but I think your arguments
> > against shared memory for web apps apply equally well to my app.
> >
> > I'll let you all know how it works out.
> >
> > Thanks for releasing it into Open Source....
> >
> > ==
> > Christopher Baus
> > http://baus.net/
> >
> >
-- 
Pavel Francirek <pavel.francirek at firma.seznam.cz>
Seznam.cz, a.s.