Status as of Friday, March 6

[Steve Sokolowski]

A lot happened this week, so it's worth an update for those who are wondering what is going on.

• Due to a snowstorm, I was stuck in Maryland for two days early in the week and had no access to the site or the Internet. During that time, something happened that caused some shares to be entered into the database multiple times. Chris eventually discovered this issue and corrected the balances of affected miners. We decided not to request repayment for one of the days that was in error, and combined with the 15% bonus on the other days, miners earned about $300 extra.
• We haven't been able to determine exactly what the cause of the bug is, but it seems to have something to do with the server losing connectivity to Bittrex. Since losses of connectivity can happen for reasons outside of our control, we needed to figure out a way to handle that situation, at least, and hopefully that will resolve the duplicate shares issue too.
• Our solution to the issue consists of several parts. First, I added a UNIQUE constraint to reject rows that are duplicate shares, so the balances of miners simply can't become incorrect due to this issue again. The constraint will cause the server to crash instead. Second, Chris created a program that will restart the mining server if it crashes every five minutes. With those two things alone, a recurrence will now result in a message being logged and mining being offline for just five minutes. Third, we expanded a logging field in the mining server that was causing exceptions because the error messages were too large to fit in the field when something bad happened. Fourth, no matter what happens, after a database operation is supposed to be done, we perform an automatic rollback instead of assuming that the user of the connection performed a commit or rollback. That will keep errors from compounding upon each other. We released all these changes in an unscheduled period of downtime yesterday evening and since then, the error has not recurred.
• Moving on to the weekend, we plan to concentrate on stability. There is an issue scheduled to be fixed that was reported in "support and feedback" about the ordering of coin addresses. We also have the actual earnings at the top of the site which can no longer run in a reasonable amount of time given the increased pool usage, so we need to figure out another way to calculate that. We need to scour the logs for any additional errors to ensure that any hidden issues are fixed. Finally, we need to fix two issues with the litecoin markets so that they can be enabled. Since we now have the pool at a state where it is profitable enough for general use, we want to focus on stability before moving on to things that can make us even more profitable.
• Please keep your feedback coming today about any issues that bother you. Since we plan to spend the weekend fixing issues, I want to find out as many problems as possible today.