I just built a MythTV box using FC5. This box is my first experience with a SATA hard drive in Linux. I installed a WD400GB SATA drive and everything installed fine, but the box locks up about once a day requiring a hard reboot. I set up a tail on /var/log/messages in an SSH window on another machine so I could catch whatever was crashing and this is what I see before the SSH connection dies:

Sep 19 14:10:29 cockholster kernel: ata1: command 0xca timeout, stat 0x50 host_stat 0x4
Sep 19 14:10:29 cockholster kernel: ata1: status=0x50 { DriveReady SeekComplete }
Sep 19 14:10:29 cockholster kernel: sda: Current: sense key: No Sense
Sep 19 14:10:29 cockholster kernel: Additional sense: No additional sense information
Sep 19 14:10:29 cockholster kernel: Info fld=0x51d3c

I am able to CTRL+ALT+F1 on the Myth box to drop to a console but cant do anything else except a hard reboot. The error messages being produced are a repeated sequence of:

Sep 19 14:10:29 cockholster kernel: ata1: command 0xca timeout, stat 0x50 host_stat 0x4
Sep 19 14:10:29 cockholster kernel: ata1: status=0x50 { DriveReady SeekComplete }

The "command" value actually varies with every error message. These messages are not being written to the /var/log/messages on the hard drive - only to the console, which suggests that the SATA interface or the SATA hard drive have crashed. The only discussion that I have seen about this are people who have had the same problem but nobody knows why it happens. The only solution that I have seen is to scrap SATA and start over with a non-SATA solution. Any ideas?

// I

I have seen these types of messages on regular hard disks (not sata). I believe they come from the hard disk controller itself - the kernel is simply translating the messages into human readable form. Generally, it means your disk is defective and probably about to fail. But look on the bright side, there could be other causes - particularly as you say it happens once per day. I would have expected a hardware fault to occur more frequently.
There is nothing wrong with SATA as far as I know, except for the connecting cables which could pick up external electromagnetic signals, disturbing the normal data flow. Try rerouting them in your box. If you have two (or more) cables, disconnect any unused ones.
You should also check the drive's temperature when you get these messages.
Whatever, this is NOT a problem with Linux as far as I can see, but purely a hardware problem. Presumably, the drive is new, so have it checked by the supplier or claim a new one under warrenty!