Performance Statistics

[Nathanial Thelen (Userplane)]

we are looking to build a very high performance system and we have been looking @ memcached.  We will be doing in the neighborhood of 200k reads per sec plus 200K sets per second.  Is this Something that with enough servers memcached can scale up to and past? If so, any idea on what kind of hardware we would need? 

Thanks, Nate 

-----Unmodified Original Message-----
It's hard to say without knowing what else is going on on that machine.
Is it reproducable enough that you can strace the server until it happens?

$ strace -ttt -s 1024 -f memcached.out memcached -u nobody -p 11211

memcached doesn't do anything in the server that'd cause it to block, so
I'm not sure what's happening to you.

- Brad


On Wed, 22 Mar 2006, Alex Greg wrote:

> Hi,
>
> We're using memcached and are generally very happy with it - it's
> serving over 2,500 requests per second on one server (Fedora Core 3,
> 2.6 kernel w/epoll), and we've generally found it to be very fast and
> a great way to offload queries from our database server.
>
> There is only one nagging problem; very occasionally (~0.1% of
> queries), we see the memcache server taking too long to send the data,
> which means the client wait for 1 second (the value of select_timeout
> in Cache::Memcached) before declaring a cache miss and getting the
> data from the database. I've reduced select_timeout to 0.2 secs to
> reduce the impact of this problem, but it still persists (albeit
> taking 0.2 seconds before declaring a cache miss).
>
> What could be causing this occasional slowness? It doesn't seem to be
> on any particular key, and seems to occur completely at random. The
> client and server machines are all running Linux 2.6 with epoll.
>
> Regards,
>
> -- Alex
>
>