So I said I would post a few more details about why the site went down on the 27th and 28th.
A little bit of background info... or: what I do from day to day
As a little bit of background, Pixel Scrapper, like every other website in the world, runs off of a nuts and bolts computer (or server), much like one you probably have at home, but a bit more powerful, and hooked up with a super-fast connection to the web so that it can support the thousands of visits it receives every day. It lives in Germany, but I manage it via an SSH connection from our home in Corvallis, Oregon. Our server is what's called an "unmanaged server", which means the people in Germany supply us with the physical computer, but everything else is up to us: installing the operating system (Linux) and all the programs necessary to serve a website, setting up all the security measures to prevent it being hacked, and monitoring it from day to day. Keeping a website like this up and running is a pretty big job on the systems administration side, and requires a decent bit of technical expertise with things like Linux, Apache, MySQL, and such. Generally you have a systems administrator to handle your server(s), and that's the person's full time job. We can't afford a systems administrator at the moment though, so I handle all of that along with fixing bugs and implementing new features on the website, and helping Marisa with support requests. As you can see, Pixel Scrapper keeps me and Marisa pretty busy, as we're both pretty much holding down what would normally be 2-3 separate full-time jobs each . We like it though, because the work is self-directed and interesting, and we really enjoy empowering others to be creative; also, you guys are just a great community to be a part of.
What went wrong
So I guess that was a pretty big tangent, but when considering website outages I think it helps to understand a bit about what goes into keeping a site up and running. What happened on the 27th was that one of the two hard drives that hold all of our server data went belly up. We're prepared for that situation, which is why we have two hard drives: each drive is an exact replica of the other, so if one of them fails, it's not that big of a deal. It does take a little while to get the drive replaced, though, and copy over all the data from the good drive, so the site has to go offline for a while for that to happen. In this case, there were also a few complications involved, and it wasn't entirely clear whether the failed hard drive was causing all the problems: some cached data became corrupted, and the site started behaving strangely. To mitigate the risk of problems down the road, I wanted to make a thorough investigation, and also fall back on a database backup that was guaranteed not to have any corrupt data hidden in a back corner. For these reasons that site was offline a bit longer than expected, and because of the database rollback, a few pieces of content that had been posted to the site were lost (a couple of layouts, a couple of forum posts, and a few comments).
As I've said in the past, Marisa and I care very much about all of the user content that is posted to this site. All of the content is backed up on multiple hard drives (we have more than the two in our server), and we take regular database snapshots to try and insure that we don't lose anything; rolling back the database is always a last resort when we face a problem. Unfortunately, in some cases, it's just not worth the risk of having corrupted content in the database somewhere, and we need to roll things back a few hours. The percentage of site content that is lost in these cases is very, very small, but we know that that fact is little comfort to the person who loses their long, thought-out forum post, or the twenty layouts they just uploaded.
We will continue to do everything we can to minimize the risk of content loss, and we hope that these database rollbacks will become even less frequent in the future. The fact remains, however, that nothing posted on this website, or any website (or saved on your computer for that matter) is 100% guaranteed not to be lost. We can guarantee it about 99.x%, but not 100%.
To minimize the risk of losing content that you work hard to create, I always recommend having multiple backups stored in multiple locations. If you are making a long and thoughtful forum post, compose your post in a word processor on your computer, and then copy and paste it once you are finished: this way you have a backup copy on your computer. Even better, compose your post in Google Docs, where it's automatically saved as you type. Before a forum post even gets saved to our website, there is a risk that your computer or browser might crash, in which case your work is lost before anyone even has a chance to see it (I'm sure many of you have experienced that). The same goes for long replies or comments. And when it comes to uploading layouts, always make sure that you have a copy of your layout on your computer (preferably on multiple hard drives): never post to Pixel Scrapper and then delete the layout from your computer. If you're uploading dozens of layouts, spread your uploads over a few days, rather than posting them all at once.
I hope these tips will help you preserve any and all of your work in the future, and I hope we can make incidents like these less and less frequent on Pixel Scrapper. Thank you always for your understanding and support as we try and keep up with all there is to do while running this website! When server issues like this come up, it puts me back several days with regards to my development plans for new features and such, and I hope you won't hold that against me.
As usual, we will be sending out extra download credits to everyone to make up for the people who missed their daily credits during the site outage: these will roll out to everyone over the next 24 hours. I've also modified our daily credit system, so that in the future it will automatically play "catch up" for anyone who missed their credits.
All the best.