Status as of Thursday, May 21
Posted: Thu May 21, 2015 9:24 am
Here's today's status:
- We successfully resolved the issue with Poloniex and brought Poloniex trading back online. Coins that are solely offered from Poloniex are back up and running. Unlike Comkort, Poloniex is very responsive and Tristan D' Augusta, Poloniex's lead developer, immediately re-enabled the account.
- To prevent this issue from recurring, we added code that will both make us more profit and make more profit for the exchanges. We will hold balances in our wallets until we have enough to exceed the minimum trade size at the target exchange, rather than send money immediately. That results in fewer transaction fees for us, and fewer inputs at the exchange. It also means that if the price at a different exchange with a lower minimum trade size rises in the meantime, we can sell at a higher price without coins being stuck in the now lower-priced exchange. We got this working immediately.
- Comkort is still not replying to our messages. We will attempt to contact their CEO today. If they do not reply, then we will take the issue to reddit and state that they are holding our customers' funds without basis and implement ccex instead.
- The universal block explorer is nearly finished. I sent it to Chris for testing. Expect to see it next week.
- User kires requested that the boxes be rearranged on the "earnings" page so that the "recent shares" be at the bottom. Since nobody has commented on this so far, we will assume that kires is speaking for the silent majority and implement the change. Anyone who disagrees has two days to say so, or the change will be finalized.
- We will be going to Micro Center this weekend to purchase a second server to ramp up our system to support multiple algorithms. While the software for multiple algorithms is not expected to be available until July, Chris will begin compiling and downloading the blockchains of x11 daemons as soon as next week.
- The block explorer may take a few weeks before all coins are available, because we need to compute a huge amount of data into the database (the table will have as many as 1b rows). Even if we dedicate eight cores to the task, it could still take several weeks to do this for the first pass, and then we'll need another day to catch up on all the blocks that were received since then. Each successive time it will take less time until we are able to add blocks to the explorer in real-time.
- This is a friendly reminder that there will likely be some brief downtime this weekend. We need to upgrade the firmware in a router to get rid of the Heartbleed bug, which requires a restart, and we need to troubleshoot the intermittent connectivity problems some users are having.