Chia farm 300TB VS 1PB - What’s wrong?

[ghost]

Finally, I got my 1PB of plots. But I do have a problem with that. The more I plot the less blocks I earn. It does not affect my earning on the pool however, daily earnings are increasing at this point.
85% of all block rewards I received with 200-500TB, and the remaining 15% was around 800-1000TB. The network at those times was always around 30EB. So, with only 400TB I received a block every 2-3 weeks, but with 1PB I got none for 45 days.
Can anyone please explain what’s wrong here? I have attached my Debug here: debug.log

1. Chia installed on Ubuntu and only farming on this machine.
2. NO errors in logs
3. 8444 port is open. I even add 300 peers now to test.
4. Farming only one more fork - Silicoin
5. All hard drives are connected directly to the farmer via USB C-type 10Gb. I use 5 ports, each has it's own controller.
6. ! I do have some external Wallets connected to my node however, not sure if that’s ok.

[loppefaaret]

there is the potential for ups a downs - although it is not common, reports of up to 10x swings have happened... a 5x swing for 1PiB is about a month, have you had 0 blocks the last 30 days with 1 PiB ? if yes, that means you are about x5 bad luck, which is going to happen to alot of people i'd imagine at some point, myself included, more than once.

[ghost]

I have 0 blocks for the past 47 days. I feel like i need to split all drives on two systems to see if anything will change.

[isipisi89]

How much energy dose your farm consume with 65 Harddrives? 650 Watt?

[ghost]

i haven't measure any power consumption from the wall, but each enclosure rated 60W,

[ghost]

I have changed configurations a bit. Now this whole wall is acting as Harvester only, and actual node is on another machine with i9 10980XE overclocked to 5GH on all cores. Will run this configuration for a week to see if anything will change in terms of rewards. Still farming on the pool.

Using 1 Gigabit ethernet switch. Unfortunately don't have 10 Gigabit switch.

[ghost]

Harvesting over local network to the farmer is a failure, i think. I have noticed ~10% drop in revenue while pooling. Feels like it's better when all drives are connected directly to the farmer via cables.