[Systems] Benchmarking aslo-proxy

Bernie Innocenti bernie at codewiz.org
Mon Nov 16 11:15:10 EST 2009


El lun, 16-11-2009 a las 03:10 -0600, David Farning escribió:
> I'll publishing the results on http://people.sugarlabs.org/dfarning/
> But, I don't understand them enough to really trust them yet.

/me reads.


> Yes, this is on my list of things to figure out.  I don't understand
> why Remora does not use APC.  My naive guess is that they feel it is
> cheaper just to add another physical machine then optimise code.

It may be that the initial versions of APC did only the bytecode caching
and could not function as an temporary object store.

Note that caches like memcached are not transparent: there's explicit
PHP code for inserting and retrieving objects in them.

Wait, I have a feeling: it may be *necessary* to share the cache among
multiple nodes for consistency. This would be the case if they store
objects in memcache to reduce the number of *insert* queries in the DB.


> Let me look at this again.  The squid cachemgr3 is reporting that it
> is handling 80% of hits out of cache.....

I bet it's just .css, .js, .png... the stuff that would be super-cheap
to serve anyway.

Squid has a command-line tool and a web tool to see what's in the cache.


> hmmm. Maybe siege is grabbing _everything_ fresh every time whereas a
> normal browser is cache stuff locally.  That would explain why squid
> is serving so many copies of *.js

Oh, that would be *very* likely. I'd make it only request php pages,
because we know already that Apache can serve 3000-4000 static files per
second, 3 orders of magnitude higher than our current limit.



> I consider you comments rather stronger than suggestions.  I consider
> them the most cost effective way to meet our infrastructure needs at
> our current scale... in terms of load, machines, and admins.
> 
> The challenge is that we are scaling rapidly.  In the long term we
> need to be able to be able to geometrically increase our load
> capacity, while keeping our machine growth linear and our admins
> constant.  This sudden increase in machines is my not so subtle way of
> pushing you (kicking and screaming) from being the Sugar Labs
> infrastructure wizard into the infrastructure team coordinator.

Then we need more sysadmins willing to take responsibility for machines,
not even more machines to be taken care of :-)

-- 
   // Bernie Innocenti - http://codewiz.org/
 \X/  Sugar Labs       - http://sugarlabs.org/



More information about the Systems mailing list