We have limited users to 1 connection in an effort to reduce the load on the fontend. We'll left this restriction as soon as the backlog has been processed.

We've turned off users to the server for a while to allow the server a chance to catch up. It's the only way to solve the missing article/completion problems. As soon as it's caught up we will turn on user access again. We've also order some hardware to beef up the server so this doesn't need to be done again. We're hoping to have this server running full speed by the weekend.

Update: In stead of continuing to fix the overview system corruption, we have opted to just erase the headers and start again. We'll have no retention for a day or two, but at least it will fix it. We also have some new hard drives ordered that should increase the speed of this machine.

The reseller admin panel now shows the wholesale costs next to the package name when viewing the list of packages.

The wholesale cost is also being computed automaticly. The pricing structure has slightly changed, but on average, it will remain the same. All packages are now priced at $0.06 per MB/day. Free free to create any size package you wish. Roll over credits are also possible by selecting the Add Method of "ADD" instead of "REPLACE".

The feeding server crashed last Friday. When the server was restored it had corrupted the counters for the article item numbers due to file system buffering. This caused it to start to number new articles with the same item number that it had previously used on other articles. This two article for the same number issue really caused problems with the frontends. So even though the feed server was brought back online with a few hours the damage it caused took takes to work out. At this point all new articles should be ok. It's just articles in the Friday - Sunday range that are an issue. I really wish we could correct it, but the only thing we can do is wait for new article to push out the old articles.

We experienced a database problem from about 5 pm on Sunday till Monday morning. The database corrupted the password list and when it was uploaded to the individual servers it prevented anyone from loggin in.