Can someone comment on my farming report

Hi to all,
Can someone more experienced give comments/suggestions on my farming report, especially on my response times which are not so great? :slight_smile:
Thanks in advance. :slight_smile:


Are your drives sleeping?
Are you sharing plots over a network?
Do your stats always look like that?
Have you added or moved any drives in last 24 hrs?
Are you plotting on your farmer?

So many things can affect response times, need that info.

Specs of farmer?
Are you using harvesters?
No of plots?


Let me answer you one by one question.

  1. I am not sure, I have no more than 20 drives on a USB connection so it is possible, I will sure check. But as I remember, the response times were similar when there were no USB drives.
  2. No, I am using 2 harvesters for no more than 1000 plots on then. Will attach them soon to the main machine. On the harvesters, I am also plotting and the response times even while plotting are way better than this.
  3. Mostly YES. If I restart the farmer or the FARMR app, the numbers are better on the first day.
  4. No. I have done some plotting but not moved/added HDDs.
  5. No, I am not. I am using at the moment a low spec CPU on the farmer but it only gets to 20-30% of CPU time so I assumed it is OK for now. I will certainly change it with I7-7700 CPU soon.
  6. Specs are: 24 Gb DDR4 RAM, low spec CPU, some Celeron G, solid motherboard, ASUS h270, plot files on drives connected to RAID cards like LSI 9264/67, LSI 9650 16M, 2 x 16 sata port PCI-E cards from Aliexpress.
  7. Yes, 2 of them, on solid Ryzen 5900 machines, on gigabit network, UBUNTU on them, WINDOWS on the farmer
  8. Around 7000

This confuses me, my setup is opposite, if I add drives or restart I get longer times in that next report, then the report after returns back to normal.

I’m to tired to think hard at the minute, need to sleep, hopefully someone else can chime in with some ideas.

What tool is being used for that output from the OP ?

Farmr ( 20 characters )

1 Like

1 - If you are using USB hubs, I’ve found plots on hubs daisy-chained to other hubs give slower response. And a high power CPU doesn’t help either, nor faster 10Mbs hubs. Seems delays are built in the the protocol under certain circumstances.

2, 5, & 7 - Your response times are better on your higher spec PCs. Not a surprise, esp. considering you are using a 2 core Celeron (you don’t say what one) on the farmer that has slow response. Your Celeron is doing ~75 lookups/min with 7000 plots. Could be some congestion there w/2 harvesters.

You say, "On the harvesters, I am also plotting and the response times even while plotting are way better than this.". Are you plotting & farming on the Ryzens as well as farming on the Celeron? And so actually have three systems farming? From your wording, it sounds like you are, rather than using a farmer-network-harvester as Chia suggests.

5 - Even thou you see only 20-30% use on the Celeron CPU, that is not the whole story of how your CPU responds to the various processes of Chia. Peak CPU usage often is significantly more than average use percent, in fact, often much more. If your CPU can’t process everything immediately, that leads to response delays. I’d guess that the i7-7700 would help a lot solving the response issues, as it’s rated several times the speed of typical Celerons. Or…

You might consider using one of those 5900’s as a plotter/farmer/harvester rather than the i7-7700. Simply delegate any required plotting to only some of its cores (set affinity to the respective Chia plotting processes.) That 5900 most certainly would cover all your needs in a single machine.

7 - Any network use (like the farmer PC, or other net traffic) can conceivably slow things down coming from the harvesters, and the reverse as well, harvester traffic can slow farmer net traffic. Perhaps not by much, but everything matters concerning network speed in getting the best response possible.


The easiest way to test remove USB ports and change to a pcie card. It will show immediately the results. I stopped all usb and only see number issues when I plot on multiple machines within my network. Once I stop all goes to normal.

1 I am using two 7 port USB hubs and have 14 drives connected directly to them. No daisy-chaining or something similar. I do recall not so long ago when I did not have USB disks connected, the response times were almost identical.
2, 5, & 7 I do not have 7000 plots on the Farmer only, it is 7000 plots on both Farmer and 2 Harvesters. The Farmer only plots are more like 5970. Interesting to mention is that Harvesters are still plotting, with CPU times close to 100% and still performing better than the Farmer with CPU times close to 10%. To me, this shows that the CPU is not a problem here. Need to mention, I am not farming on Harvesters, I am using them as proper Harvesters connected to the Farmer machine:

I will certainly replace the Celeron CPU and report here what is the case. But I highly suspect the response times will go down.
I can not use one of my 5900 machines for Farming. My config is specific, I have 8 PCI-E slots on my current MOBO and all of them are full with RAID/SATA cards. The mobo on the 5900 CPU can not provide the same functionality. Plus I might sell the 5900’s once I finish plotting.