I would like to clarify some issues regarding the missing hand histories and ask some potentially relevant questions in response to UB's official statement. If you are interested in my background, I am a casual amateur poker player but I have been a network administrator for 10+ years and have worked with a variety of RAID server configurations, database management suites, and enterprise backup software packages.
http://blog.ub.com/2010/11/hand-histories/
1. The relevant hand histories originated from the Excapsa Software package (it handles all the real-time data analysis in memory as hands are being played live.)
2. After hands are played, the data is sent to a database to be permanently stored on a hard drive.
Q: Was the database storage solution created by and part of the Excapsa Software package, or was it a publicly available database management software package such as Oracle or SQL?
3. The database was stored on a server with RAID configured hard drives.
Q: What type of RAID configuration was used on the server (assuming it was not RAID 0 as this is not really making use of the RAID method and there would be no purpose in labeling it RAID)?
4. One of the RAID drives failed.
Q: Was the array rebuilt by replacing the failed drive with an identical functioning drive? If so, what errors were encountered in trying to rebuild the array? Was the RAID controller and/or the RAID configuration damaged? If the controller was not damaged then the configuration would still be healthy and the data can be recovered (in most cases automatically) by replacing the failed hard drive with an identical functioning drive. This is the purpose of using RAID.
Q: Was the damaged disk sent to a data recovery specialist? Was the damaged disk destroyed or reused? Is UB willing to publicly submit the failed hard drive for testing to a respected data recovery company? Data recovery specialists can re-create the data on a hardware level by reading the individual magnetized sectors on a physically damaged disk and can even emulate RAID array configurations to re-create the data.
5. Backups were “lacking.”
Q: Were backups made of the database in question at any point in time? Were there any backups of the RAID configuration? It is unclear from the post whether backups were never made or were lost. At face value, this seems to be an absurdity caused by negligible incompetence, at best.
I have been involved in data loss situations where the RAID controller malfunctioned or its configuration was accidentally erased and we didn’t have a backup of either the RAID configuration or the data, but NEVER from one hard drive failing in a RAID array.
Either the array can be rebuilt, which is usually the first option available in the RAID controller's software interface in the event of a drive failure, or the drive's data can be physically extracted and rebuilt by data recovery specialists.
I would also like to comment briefly on the issue another poster alluded to with (part of the?) cheating program being eventually discovered as a localized set of registry entries on the cheater's machine. If this is all that was required to see another person's hole cards, and the program was eventually hijacked and distributed by person's not in the original conspiracy, how can we know that the cheating was ever limited in scope to a certain number of specific accounts that did not understand how to be cheat discretely in poker terms? Did the server only broadcast hold cards to specific accounts that were able to receive the info through the registry entries, or could anyone with any account be a superuser if they had the registry entries? Is this something that UB has or will investigate (do you care?)
Thanks for taking the time...