Compensation for Missing Shares

Yesterday morning, we deployed new code for our mining server that dramatically reduces CPU usage and resolves some minor bugs. However, there still are bugs in the new code, and the mining server stopped recording shares around 9:30 AM. I discovered the issue around 1:00 PM and reverted the mining server's code, but for those four hours no one earned any money despite submitting shares.

I will be crediting users for the lost mining time based upon hashrate and average earnings for the four hours before the lost shares. The credits will be applied toward today's balances, which will be paid as normal tomorrow. I will make another post once I have made the corrections. Thanks for your understanding.

The additional money has been added to everyone's accounts. If you were mining between 5:53 AM and 9:53 AM yesterday, you received 4 hours of extra earnings to compensate for the missing shares.

The mining server may restart a few times over the next few mornings while we work out bugs, causing short disconnections. I will be closely monitoring the system to ensure that there is no additional downtime due to a crash.

The mining server crashed again on the morning of August 25, for a completely different reason. We are in the process of modifying the database to improve the response time of the block explorer, and that requires us to update about 2 billion rows. Today, one of the queries we are using caused the database to run out of memory and crash. The mining server is supposed to cache shares when the database is offline, which it did. However, the mining server ran out of memory to cache shares before the database outage was noticed and crashed itself, which caused a loss of shares.

Just over three hours of shares were lost between 10:30AM and 1:30 PM. I have credited all users for those missing three hours based upon the earnings from 7:30 AM to 10:30 AM. The earnings have already been added to your balances.

I don't think we will have another one of these issues again. I have modified the query to update fewer rows at once, reducing memory usage on the database. The mining server has had its bugs fixed on Wednesday, so that shouldn't cause crashes either.