I hope someone can help me here. I started with my little chia farm with 3 Win10 systems that all ran as Full node. And in the beginning everything seemed ok. My look up times were always below 5 seconds when I checked my log files and at about 800 plots I won my first XCH so I thought all is just fine.
Farm grew to about 2000 plots now and I switched to MadMax Plotter meanwhile. Checking log files again, I was surprised to see that my look up times became pretty bad. Average between 10 and 25 seconds and not a single one even close to or below 5 seconds. So I changed the setup: Got rid of three full nodes and am now running 1x Farmer and 2x harvester via harvester protocol. It improved look up times but not enough. I reduced Threads for madmax (which was running on all systems) until my CPU max activity was about 50%. Then I switched plotting off completely, CPU idle. But still not a single look up time is below 5 seconds. What am I missing here? Please can anyone help me?
My configuration:
Farmer: Dell T5810 Workstation 14 core, 64GB, 3xHDD connected, running chia GUI.
Harvester 1: Ryzen 7 5800x on ROG Strix B550-f with 64GB and Adaptec 71605 HBA with 10 SATA drives connected internally.
Harvester 2: Identical to Harvester 1 but LSI 9217-8i HBA card instead with 10 Sata and Sas drives connected internally.
No usb drives. No nas. No drives connected through network. Port 8447 is open on farmer. Upnp is active. Network is all cable bound, gigabit network. Internet connection through cable. Can’t open port 8444 somehow in my router. But this was not a problem earlier, when I won my first chia.
I hope, I didn’t miss anything. Really hoping anyone can help here!!
Got the same issues with R510, tried all setting to improve the response time but no hope.
Then i Stoped using madmax from the harvester and all go to normal.
But he’s saying he already stopped plotting and still has too long times.
I would try to shut down both harvesters and see if just one machine will run ok or not.
If not then most likely some sort of network or system issue has come up.
Unless you can think of things that you might have changed or other thing that might cause this, I would reinstall the OS and Chia, maybe even reset the router and input new settings.
Strange that you cannot open 8444, you might consider running a vpn that has the port-forwarding option in it?
Clearly a resource issue, likely due to configuration. What is that? I don’t know. Not enough detail here. Other thing I’d offer which is also not too helpful given the level of detail that’s not in this post, you definitely don’t need 3 systems to farm 2K plots. Nothing wrong with USB and network (NFS) drives if you know what you’re doing.
I did turn everything off and on again. I also stopped madmax. I even stopped plotting in general. But lookup times stayed above 5s. I have run perfmon on windows to check IO of my disks but that’s all within normal parameters.
employ harvesters. don’t under any circumstances. use nfs cifs or anything like that. keep your network pure… the farmer doesn’t need to be a harvester… so run remote harvesters. than turn off your farmers harvester. dont plot on the machine your farmer is on.dont plot to final directory over network. only rout network so the harvesters can talk to farmer.
and dont put your full node on a pie
peace
You may have done this, but didn’t see that in a quick read through.
Given all you have done, which is a lot of troubleshooting on your hardware, a suggested next step is deep dive into your network. You could start doing some investigation on web sites giving tools looking at all net aspects like > 9 Best Network Bandwidth Monitors (Free and Paid) - DNSstuff.
Another avenue to explore is your internal network itself. What cables are you using, are they pre-made or constructed. Have you an analyzer to test them? Check for ethernet errors, check for bad ethernet adapters, check for old or failing switches/routers.
Hook up a simple laptop after disconnecting everything else at your router and do some analysis at that point of goodness of your external connection (to remove all possible other blockages/network storms/etc).
Basically, divide and conquer. Remove as much from the chain as possible and work backward to find problems. Good luck!!