Status as of Sunday, April 8, 2018

Good morning!

• All maintenance was completed successfully at around 5:30am EDT this morning. Chris went to sleep and then will return to State College after he wakes up. Once he arrives, he will fix a few remaining minor software issues, which could increase orphan rates slightly until then.
• The network connection is fully working. We now have eight mining servers, up from 4 before. Each of the mining servers is running on faster cores, with Chris having resolved a "power saving" mode in their BIOSs. The mining servers are all using the "invalid authentication" bans that have resulted in improved performance. We did not see any cases where the CPUs reached 100% load during the hour I watched them. Chris also replaced and trashed an overheating motherboard in one of the coin daemon servers.
• I was able to finalize part of the low-luck miners investigation and account for 3.2% of the missing luck and develop a 100% solution to completely eliminate these losses. I'll be writing a detailed description of the problem as soon as I have time to do so. That means that about 6% of the missing luck still remains to be accounted for. I want to let the system perform today, now that the code has been applied, and reevaluate the luck situation as it stands now.
• My main focus today will be resolving support tickets and preparing lessons for Vance Vu, who is driving to State College today to begin his new position tomorrow. Please note that we are still in reduced support mode. There are a lot of very complex tickets, mostly related to payout issues, which can only be resolved with Chris's experience. He won't be able to look into them until at least tomorrow evening, because he needs to drive this evening, finally get a full night's sleep, and train Vance the next morning. I apologize for the delays; once Vance is fully trained, the purpose of his hiring is to permanently resolve some of the backend issues that cause the tickets, particularly issue involving payouts.

so its safe to move the miners back to prohashing?

pumaro wrote:so its safe to move the miners back to prohashing?

The level of safety you require is up to you to decide. In my opinion, the level of service is now sufficient such that all miners can return. We have capacity for around 45,000.

After a massive upgrade like this, the discovery of bugs follows an exponentially declining function. In the first hour, there will be a lot of bugs, and by the second hour, only half as many, and by the third, maybe a third as many, and so on. By now, the bugs we are aware of are down to the very minor ones - a connection limit to the website, a misconfiguration that causes increased orphan rates on one daemon server, and a few others. The number of bugs will never reach zero, so it's up to you to decide when a good time to return is. I'm using the system to mine enough to heat my basement with no issues, so I'm confident enough in it.

Even immediately after things came back online, I think the system was already at the point where the upgrades themselves had added so much CPU power and reduced latency so much that the initial bugs were outweighed by the improvements that the upgrades caused.

since u have resolved the luck situation, I assume those of us who received any payment in past few days will now start to receive what we are owned ?

cheers

A big call out to Steve and Chris. Thank you very much for your hard work!

waynefatboy wrote:since u have resolved the luck situation, I assume those of us who received any payment in past few days will now start to receive what we are owned ?

cheers

The luck situation hasn't been resolved. There is still a problem. What I did uncover was a scam that was limited to static coin miners only that reduces the scope of the low luck problem to about 6% from the 10-11% I originally thought it involved. That will be explained in a different post. The number of low luck miners was always identified to be around 6%, so that's why I spent the last week trying to determine why so few miners were being identified. It turns out that the cause of the difference was a scam, so we are making progress now that we know those involved in the intentional scam are a different class of people who we will pursue in parallel while we continue the investigation into whether low luck mining is intentional or not.

On another note, I wanted to point out that the difference in orphan rates has been dramatic, far better than I predicted. On some coins, we've been seeing orphan rates down from 17% to 5%! Our calculations originally were that lower orphan rates could at best earn us $700 per month, not enough to justify the $3500 cost of the connection on their own. So far, it looks like the lower orphan rates may pay for the upgrade by themselves.

I think your fixing the "bad luck" has not solved anything. I put my machine back to run another 24 hours and NOTHING. I have a balance of $14.50 with bitcoins and another $2.5 in Litecoins.

I want the payment to be made or please at least let me know if you are going to be honest and pay me what my machine earned or let me know that you are NOT going to pay and I will simply disappear and never be back here again.

This is my 10th message and no answer, so, is this pool a real pool or another scam?

Thanks. Fausto.

Fausto wrote:I think your fixing the "bad luck" has not solved anything. I put my machine back to run another 24 hours and NOTHING. I have a balance of $14.50 with bitcoins and another $2.5 in Litecoins.

I want the payment to be made or please at least let me know if you are going to be honest and pay me what my machine earned or let me know that you are NOT going to pay and I will simply disappear and never be back here again.

This is my 10th message and no answer, so, is this pool a real pool or another scam?

Thanks. Fausto.

There has not been a fix for bad luck. There was a fix for static coin miners, which I detailed in the previous post.

The low luck issue is still under investigation. There is definitely a problem here; we have never had a day where the number of Digibyte blocks, for example, has been above the expected value, out of the past 60 days.

There are also some customers who are extraordinarily unlikely to have been misidentified. I ask for your patience while we continue to investigate the issue. If you have to use another pool in the meantime, then I think that's is your best option for now.

Fausto wrote:I think your fixing the "bad luck" has not solved anything. I put my machine back to run another 24 hours and NOTHING. I have a balance of $14.50 with bitcoins and another $2.5 in Litecoins.

If your total balance is $14.50 with bitcoins, don't expect to get a payout.
You need like $100+ minimum to get a payout with the threshold on Bitcoins.

Steve Sokolowski wrote:There is definitely a problem here; we have never had a day where the number of Digibyte blocks, for example, has been above the expected value, out of the past 60 days.

Steve, which Digibyte are you referring to? the Scrypt coin or the SHA-256 coin.
Are the mining numbers being combined across algos? Wasn't the SHA-256 just recently added in the past month or so?
If you look at JUST the scrypt digibyte does it ever go above the expected in the last months?


Just pointing out, digibytes is a coin that exists on 2 algos, if that has any impact or not.

CSZiggy wrote:

Fausto wrote:I think your fixing the "bad luck" has not solved anything. I put my machine back to run another 24 hours and NOTHING. I have a balance of $14.50 with bitcoins and another $2.5 in Litecoins.

If your total balance is $14.50 with bitcoins, don't expect to get a payout.
You need like $100+ minimum to get a payout with the threshold on Bitcoins.

Steve Sokolowski wrote:There is definitely a problem here; we have never had a day where the number of Digibyte blocks, for example, has been above the expected value, out of the past 60 days.

Steve, which Digibyte are you referring to? the Scrypt coin or the SHA-256 coin.
Are the mining numbers being combined across algos? Wasn't the SHA-256 just recently added in the past month or so?
If you look at JUST the scrypt digibyte does it ever go above the expected in the last months?


Just pointing out, digibytes is a coin that exists on 2 algos, if that has any impact or not.

Yep, I knew that, but thanks for the tip. That would make sense, but $PAC is also consistently below the expected value too, with only 1 day in 60 making back the same money paid out to customers.