Our machines didn’t like us yesterday (maybe we worked them too hard). The site was slow, and later that evening, part of the database cluster broke. The site was never offline. However, some people noticed that after they clicked on a buddy to reset the unread message count (the number next to the buddy), the number didn’t reset. Others modified their account settings, then clicked “save”, but the changes didn’t take effect. Why?
We have several databases that are copies of each other. In case that one computer breaks, the others can pick up the slack. In other words, whenever you made a change in Spokeo, that change is saved multiple times on multiple computers to guarantee data security. This process is called “database replication”.
The database replication broke yesterday, so the data wasn’t copied correctly. This is why you saw inconsistent behaviors on Spokeo. Fortunately, we noticed this problem within hours, and the problem was fixed last night. Thanks to database replication, no data is lost.
We had encountered all sorts of problems and bugs before, and we will have more of them in the future. After all, computers break, and softwares crash. Other than trying our best to fix these bugs, we will keep you updated with all our activities on this blog. From now on, we are going to make this site-building process as open and transparent as possible. We cordially invite you to give us feedbacks on how to improve further.
On the last note, we will be releasing a small upgrade to our system on Wednesday. The site will be down for couple hours, but you’ll appreciate the new changes.
