memcached subscription notification

Brad Fitzpatrick
Tue, 25 Nov 2003 12:24:02 -0800 (PST)

Servers don't know about each other at all.  Memcached is a hash of
hashes:  the client does the first hashing (a key maps to a certain
server), and the next hash layer is done within the server.

You'd run multiple processes on a 32 bit machine with more than 4GB of
memory.  Each process can only see 4GB, but you might have 64GB of memory
in a 32 bit machine with PAE.

- Brad

On Tue, 25 Nov 2003, Larry Leszczynski wrote:

> On Tue, 25 Nov 2003, Brad Fitzpatrick wrote:
> > Feel free to post to the list if you have questions.... everybody's
> > friendly and helpful.  :)
> Thanks, Brad.  Now that you mention it, I do have some FAQish questions
> that I didn't see answers to after a quick scan of the web site and list
> archives.  Sorry if they're remedial but I just learned about memcache (on
> the mod_perl list) and wanted to get a basic idea of what it does and how
> it works:
> 1) It looks like the client initializes with a list of servers it knows
> about, how do those servers communicate with each other (assuming they do)
> so they stay consistent with each other?
> 2) Assuming the servers have some mechanism to stay in sync, is there a
> way to clear the cache across all of them and start from a clean
> state?
> 3) I can see why you'd run memcache daemons on multiple machines, but what
> would be the motivation for running more than one daemon on a single
> machine?
> Thanks,
> Larry Leszczynski