Status as of Tuesday, October 18

[Steve Sokolowski]

Here's today's status:

• The database upgrade was completed successfully. Chris ran into an interesting problem. The hypervisor containing the database server had not been restarted for 1.6 years, and apparently a cosmic ray had caused a bit to flip in the "cp" program. When he tried to copy 1.1TB of upgraded data from the hard disks back to the database solid state RAID, there were kernel panics. It took Chris some time to figure out what had happened and restart the hypervisor, but everything is fixed now. He also thinks that the system will be more stable and have fewer random database crashes. I'm going to suggest to Chris that we perform restarts every month or every 3 months to limit the possibility of errors building up in memory over time.
• We are aware that there are a few errors with the website. In particular, there is a display error with the black bar at the top, and some miners are stuck at "calculating hashrate." We expect to have these issues fixed before the weekend. These issues are not affecting profitability and can be ignored for now.
• The newest release of the mining server, which will be the final single-algorithm version unless there are major bugs, was released after the upgrade. This version corrects a misunderstanding of the mathematics involved in work restart losses due to stale merge-mined work. The newer version sends work restarts less frequently due to merge mining and will improve profitability for all miners. The slowest Titans may see improvements of up to 5%.

[ltc2084]

You guys rock. Hoping to switch back soon if I can get that KnC firmware.