I completely agree that mining is a hobby. The relationship between miners, their hardware, and popular coins, however, is much more complex. Just look at LTC vs ETH, where LTC was created as a GPU-miner's safe haven against BTC's ASIC revolution. Once scrypt ASIC's put GPU-LTC mining out of business, it faded. Now the new, popular coins (ETH, Zcash) are the GPU-friendly coins again. Entering crypto with a mining rig is a great way to support the network, learn something about crypto, and take some risk. I can agree with OP that individuals should not go forward with mining as a profit-seeking endeavor.

I have been paying 9-12 c/kwh since 2013. In the winter months (nov-mar) it is 0 c/kwh since my home is heated with a giant resistor, and this offsets that.

GPU mining was dead in a few months for how many years now? I put together a 4x 7950 rig in 2013, and it's been more or less profitable for 4+ years despite bitcoin, scrypt, and others all going ASIC. It's not always highly profitable, but at a minimum I get free power for 3-4 months a year because I account the heat output as equivalent to my resistance-based heater. By the time the residual value of the hardware is calculated, it's basically a medium risk hobby.

The reality is too many people want mining to be GPU-based for the ecosystem of mine-able coins to disappear. CPU-based lends itself too easily toward botnets, ASIC based lends itself to centralized manufacturers with the NRE capital, and that leaves GPU as the quasi-distributed platform that enthusiasts center around.

Erundook you must relocate one of your pools, close one, or do something; you have too much hashing power connected on the same subnet. Forking has already happened and will happen again. Please recompile your worldcoind's with outgoing connections over 8 (pref >250) and incoming over 125 (pref >250) if you haven't done so already.

Block 41306 to 41670 got orphaned on the chain, but it wasn't just my own fork. My pool got 205 blocks out of 364 which means someone else is out there on a forked chain.

We absolutely must move some hashing away from the eurndook pools, and not to my pool. 650+480=1130 MH/s sitting on the same subnet in a datacenter, so this is going to happen again.

Well maybe, but it wasn't the only node contributing blocks to the chain. It looks like the "erundook" nodes combined represent the 1200 MH/s that was on one fork, and my pool had ~60% on the second.

Was there a fork just now?

Successful Trade

Hosting provider (Staminus.net) outage impacting dontmine.me

You do realize a second server by the same pool operator is not the same thing as spreading the hash rate around right?

Booters cost like $5/month and the pools need corporate grade protection to counter them; it's not cheap (2-4k/month for TCP applications like Stratum). If we manage to knock off 50% of the nodes, the booter price might go to $10/month or something so it's still a losing battle. These aren't the sophisticated attacks that mtgox has to deal with, but a simple UDP flood. Most hosts that offer DDoS protection, from my shopping experience, max out at 10Gbit/1-5MPPS, and I consistently saw attacks stronger than that with CNC (peak was 22Gbit, 75% of the attacks were over 10).

Some prices for dedicated DDoS protection I found: (not shared like awknet or VPS)
Staminus     $1k/month for 10Gbit/1MPPS (not strong enough)
BlackLotus $675/month for 10Gbit/6MPPS
Some other  $1k/month + $4k setup for similar

The solution I've come up with is to just use a suite of reverse proxies:
buyvm/etc VPS (10Gbit/5MPPS)
Minecraft-oriented VPS/Dedicated (Varies)
Cloud Load Balancer

For example, I used an amazon elastic load balancer and some micro instances for forwarding. By using the ELB, amazon soaks up the packet floods and does some filtering. I also use cloudflare free, but there's a risk. If your site gets a http-layer attack, and you're not on the 200/m plan, cloudflare will change your DNS record and effectively direct the traffic to your server. The pro is a packet flood goes to the CDN node, and that's not associated with any single domain, so it blocks those (you can route longpolling through cloudflare since it's HTTP traffic)

Check it out now, the new ELB setup was messing with the IP addresses.

./worldcoind listtransactions "" 10000 | grep orphan | wc -l

./worldcoind listtransactions "" 10000 | grep generate | wc -l

./worldcoind listtransactions "" 10000 | grep immature | wc -l

Gives you counts

Pool Efficiency: 89.52% , about 10% stale/reject from the miners, pretty good considering a 1 second latency on a 10 second block is 10%.

WDC Network Latency + Poolserver Workgen Latency + Pool<-->Miner Latency + Miner CPU to GPU Latency (Can be large with high intensity)


Edit: I'm not trying to be showoff or anything, just suggesting that it's possible to do with better configs.

I'm at 3.5% orphan rate on my pool.

Any DDoSing yet? I haven't noticed any but it could just be blocked by the new setup I'm using.

I made it easier for you.


Hash rate should be better now. It's just a calculation, has no impact on payouts (only shares do).

You can test this:

http://wdc.dontmine.me

Probably some bugs in there, we'll see.

I'm using stratum locally and it's not looking good with my single miner, no point to pool yet. Blocks are too fast.

Do I dare setup a pool for this?