Hello all, Alex here.

First off, let me say happy new year to everyone! I really enjoyed my holidays this year, and I hope you did too.

That said, I'd like to make a new years resolution of sorts. Just prior to Christmas, our "recode" version of wardrobe was opened to Gold Account holders for beta testing, and will probably officially replace the old wardrobe before the end of January.

The Recode is a total replacement of Subeta's programming and is absolutely vital to ensuring that the site can continue to grow. The problem is that we're already feeling these growing pains. Recode will eventually fix these, but only when a substantial part of the site is running recode programming. I don't want to wait that long, and I'm sure you don't either.So I'm beginning a "stability project" for Subeta, of which recode is only a part. The stability project will address issues in Subeta's code, servers, database, and handling of downtime. The goal: 99.9% uptime/month by Dec 31st, 2009. I think we can make it!I won't go into much detail about the project here, but I will list some key points:

* Automatic recovery from SQL lockups: This is actually already working. You may have noticed (I hope you have) that the site doesn't go down for hours at a time now. The lockups are now detected and the database reset automatically, within minutes of it happening. No more waiting for Keith or myself to notice the problem and correct it.

* Tune-up of Subeta's old code to reduce the load on the database and eliminate "bottleneck" points that cause lockups.

* Changes to old code necessary to support running the site on a MySQL master/slave system instead of on just one database server.

* Downtime tracking and reporting! We'll need to keep track of when and how long the site is down to measure progress.

* And many more small changes that I hope will make a big difference!

None of this can happen overnight, but expect to see things gradually improve over the course of 2009, starting soon! Thanks so much for being patient with us in the meantime!I hope that was simple and understandable enough.


I'll address a couple of FAQs:

Will this delay the recode (much)?

Yes, but like I said, better to fix the problems now rather than wait for recode to fix them later. It shouldn't delay it by much though, recode is part of the stability project and ultimately the goals are the same.

That sounds nice, how can I help?

I may have need of some testing volunteers eventually, but for now please just be patient with us.

Why haven't you done this sooner?!

Really, it is a work in progress. Up until now though, there really has been no yardstick for progress made, other than "we had a lot of database problems this month," etc. The idea of the stability project is to attain a stability goal (99.9% uptime/month) by a deadline. With the ability to measure downtime, we will be able to focus our efforts on those activities that are having the greatest effect, which often are not the ones you'd think would.

Does this mean you will be fixing the ______ bug?

Most likely, current bugs will be fixed when their area of the site is recoded, unless they are related to stability problems. As it is, it is hard to tell whether bugs are caused by bad coding or simply what I call "server weather conditions" (e.g. high load causes a mysql connection reset in the middle of an attempted trade), which are more difficult to track down.

As always, we really appreciate your patience and support. Without you guys, we wouldn't have anyone to do all this work FOR!