Analog Reports are Back Up

Okay! Reports are working again. Thanks for your patience, everyone.

This is a much simpler design, and so far, seems to be a lot more stable. I should be able to start adding new features to the conrol panel (like actual controls) very soon.

Again, the traffic reports are at:

https://secure.sabren.com/cornerhost/controls/

Login with your normal username and password.

Control Panel Update...

Almost done fixing the communications layer to get the reports working again. Sorry for the delay - it was a harder job than I anticipated.

The control panel talks to the individual servers (mercury, scandium, and soon titanium) by means of a secure connection to a custom XML-RPC server written in python. For whatever reason, the SSL wrapper I was using was causing all those problems, and there wasn't an easy way to write a secure XML-RPC server in python. So I had to write one myself, using bits and pieces of other modules. I just got it working (it's 2:00am)... I'm going to bed, but the control panel will definitely be back up tomorrow.

Anyone Lose a MySQL connection to a private server?

Over the past couple days, *incoming* traffic on Mercury shot through the roof. 5-15 gigs of bandwdith were going into the server every day. Rackspace wasn't a lot of help, so I had to crack open the books and figure out how to diagnose individual network packets.

I discovered that all the traffic was coming from one single machine, but was unable to see the actual data being transferred . I interpreted that as some sort of denial of service attack, and blocked all traffic from it.

Now that I've had some time to think it over, I realize that someone was actually making a mysql connection to this other machine and running hundreds of queries.

So. I've probably broken somebody's website, and they're probably wondering what the heck happened.

If you're that person, please get in touch with me so we can figure out what to do with this. :)

scandium back up

Rackspace has rebooted scandium. It should be working fine now.

scandium reboot, control panel down

The trouble that hit mercury back on the 11th (extremely heavy load, unable to see processlist, etc) has happened again on scandium. It seems to be related to an unsolved red hat bug.

As far as I can tell, it's caused by one of the programs I'm using to let the control panel communicate with the individual servers... So, I need to reboot scandium and also take down the control panel temporarily. I will try to get the new version working as quickly as possible - hopefully before tonight.

mercury troubles

Rackspace is helping me track down some extremely heavy incoming traffic on Mercury. Some services may break temporarily...