today, i spent working on a project that will not produce anything, will not make anything better, will not amount to anything… and i feel *fabulous* about it.

i spent today working on setting up a test for the puremessage servers, working on throwing mail at one of the milters as fast as i can to try and make it choke. i actually got to run one example test — not a valid test, but sort of a proof-of-concept that the structure of the test is going to work — yesterday, but i didn’t get all the ducks in a row to actually run one of the actual tests until late this afternoon.

disappointingly, our fastest development machine — an apple xserve g4 with an xraid appliance — can only throw about half the traffic i need to generate (roughly 100k messages/hour, or about 1700 messages/min) at the milter, so i’m going to have to move on to other machines as well. i have had absolutely *fantastic* luck with the xserve — and it’s a g4, not a g5! — it blows our toy suns out of the water, absolutely. it’s the fastest goddamn machine, shit you not. so i was hoping it would be all that i needed for that particular portion of the test, and i also don’t know what i’m going to have to dig up in terms of other machines to generate spew. but that’s for tomorrow.

right now i’m generating about 650 messages/min, and the milter host is holding up relatively well — under its suboptimal configuration (the one we used on rollout day, which we had to back out because its performance was so awful — but that was with 8GB of core, and now these machines have 32GB). so i’m very hopeful about the optimized configuration that we got from their pro services. if i can handle the load with the crappy configuration, the non-crappy configuration should be way better. at least, that’s my expectation, and hope. i may work late tonight just to find out, because i really want to know. and i really want something to go well. if this goes well, then my life will look much more pleasant going into my trip to lisa.

of course, if it doesn’t, i have some serious problems… not the least of which is “why the hell isn’t this helping when anyone would think that it would” …but that is a worry for later.