Quote from: kano on October 25, 2017, 12:53:59 PM

At the moment I'm still trying to get stats from certain large miners to find a single case of a miner finding a block that didn't show up on the pool.
This would show something wrong other than bad luck.
Unfortunately a lot of the big miners reset every so often and thus lose all the stats up to the reset point, so I've not been able to check that many.

Anyone else here of course if you have time to do it and can, please check your miners.
What I'm looking for is a miner with a "Best Share" above Network Difficulty (1196792694098.793)
Any shares found that should be blocks, I can convert the "Best Share" share difficulty into a block hash and check it matches one of the blocks on the pool.
There aren't hundreds of blocks in the last month, so it's not hard to check.

Meanwhile, if this continues on for much longer, I'll consider putting the pool into maintenance mode for a few days (i.e. stop all mining) and continue my investigation that way.
I'm considering going back to the old code I ran until segwit required the update, and re-implementing segwit and none of the other random changes in ckpool.
I've been running the same code non-stop since August-8 when I added the 🐈 to the coinbase, and of course we've not had anything like this before this week.
The Aug-8 change was only a minor change to the earlier Jul-20 change for segwit. There's been no other changes I've implemented, in the past year, to the mining code.
The most difficult issue is the fact that we do find blocks, rather than none at all, thus if there was a problem, it wouldn't be as easy to reproduce.

Some large miners have already moved hash rate away, temporarily, due to the cost issue of paying very large electricity bills and other related costs.

I'm also considering completely blocking NiceHash, for any who use that, since their stats are pretty close to the 'unexpected' limit at the moment

I have drawn the conclusion the slowdown started in late July
I also do not rent much from nicehash as my luck is not very good with them. Since most hash was pointed at mmpool where hitting the block means you get the tx fees I keep track of the blocks I hit with nicehash.
I hit a few lower then I should have. So my conclusions kind of mirror kano's. I am okay at a few things no where am I in Kano's league as a coder.