Hi everyone, now i’m farming chia plot by USB port (32-34 disk/pc) with 4-5 hub usb (7-10 port 3.0) .
Sometime, one or a few disk crash 100% (Task manager) → all the other disk will crash too → this PC lost farming point.
So i need a tool or script can detect and notice when a disk crash 100%.
thank you!
I had a similar problem, turned out to be a problem with the USB hub, without it everything runs fine now.
Maybe just try another hub?
The error I was getting in windows was “raid port reset”
I don’t know of any script that can do what I think you want, maybe somebody else will.
without further details (error messages, screenshots, steps to resolve the issue, nothing can be done anyways.
Couldn’t he run a script that attempts to write a few bytes to each drive every 60 seconds or so, and upon a failure, an e-mail message (or similar) be sent to his smart-phone or the device of his choosing?
Or how about a script that scrapes relevant information from “chia farm summary”, and when the number of plots goes down, he gets notified?
My script writing skills are not up to such a job. But I believe one could be made by someone with good script writing skills.
By the way, like @Voodoo, I too had a “hub” related issue.
It was a Sabrent 16 port hub. After I removed it, my external drive issues were no more.
sth like this you mean?
chiawatchdog.ps1
# Declare variables for email recipient, SMTP server, password, max plot count, and threshold percentage
$emailAddress = "example@email.com"
$smtpServer = "smtp.example.com"
$password = "your_password"
$thresholdPercentage = 0.9
# do not change this variable, it updates automatically
$maxPlotCount = 0
# Start an infinite loop
while ($true) {
# Get output from chia farm summary command
$output = chia farm summary
# Check if output contains a plot count
if ($output -match "Plot count for all harvesters: (\d+)") {
$plotCount = $matches[1]
# Check if current plot count is higher than max plot count
if ($plotCount -gt $maxPlotCount) {
$maxPlotCount = $plotCount
}
# Check if current plot count is below threshold percentage of max plot count
if ($plotCount -lt $maxPlotCount * $thresholdPercentage) {
# send warning mail if treshold is reached
Send-MailMessage -To $emailAddress -Subject "Low Plot Count Warning" -Body "The current plot count is below $($thresholdPercentage*100)% of the maximum threshold of $maxPlotCount. Current plot count is $plotCount" -SmtpServer $smtpServer -Credential (New-Object -TypeName System.Management.Automation.PSCredential -ArgumentList $emailAddress, $(ConvertTo-SecureString -String $password -AsPlainText -Force))
}
}
# Sleep for 10 minutes before checking again
Start-Sleep -Seconds 600
}
the smtp/password stuff is untested, maybe needs minor change. Its just out of the blue. I dont have such issues so far.
If you run smartd, it will notify you about missing drive. I used to run expanders with risers, and whenever expander crashed…i have got mail
this tool for linux? now i run Windows, do you have any name like smartd on win?
if you do Windows, don`t bother to do anything serious 
smartd is most likely ported to Windows as well, but there is missing whole UNIX kernel so no idea whether it will work. It uses cron for checks, and mail server.
If you like challenge, there is cygwin or Win10 has got WSL2 nowadays where you can install Ubuntu in VM.
I would suggest you to learn Linux such as Debian…you will have 99% less problems like now 
My Linux servers running for 20 years, and 99% of problems were between chair and keyboard.
Of course, if you “losing” drives in OS…that is HW issue, and Linux won’t resolve that…by it will at least tell you via email.
I disagree. Ive had windows servers for just the same time as well without a single reboot.
my windows server is up for 161 days as of right now, no single issue:
In fact I had more fatal issues with ubuntu so far. One single typo, and your system is garbage.
Nvidia drivers blocking your whole package management, etc.
I have ubuntu servers running fine now. But boy, what a pain that was. I would definitely not say that Linux is more stable per see.
One single typo, and your system is garbage.
there you go, infamous chair/keyboard syndrome
i do not like Win platform for anything serious because of MS Self-Fuck TM system…every update…restart, and P&P (Plug & Pray). Windows is like modern princesses…high maintenance, little or no real use.
of course, there is not bad or good…just badly used.
why the hell you people still run plotter on Windows is beyond my economical understanding…you lose instantly at least 20%…not to mention RAM disk is crap, instable.
I am guessing from that screenshot, you are running PBO 2 on 5950X…it uses +30% more energy while giving you 1% more plots/day than using manual OC. I run same CPU with 4.3GHz @ 1.15V…170Wh or 52W/plot.
The reason why the hell my people and I run our plotters on Windows is because those are the boxes that we already have, and also because Windows is the OS that we are familiar with.
Also, when I (and perhaps my people) started with Chia, it was our first venture into digital currency. I did not want to complicate matters by installing an OS that I would have to learn. If a problem arose, then troubleshooting whether it was a Chia issue or an OS issue would have made things very difficult.
In my case, I started off with a Gigabyte motherboard (the only one I could get my hands on during the shortage), and my USB drives constantly disconnected. It turned out to be a BIOS issue, which Gigabyte acknowledged, released new versions, and yet the issue persisted – and they are impossible to contact in any way.
Had I been using Ubuntu, I never would have nailed down the problem.
Someone that has Linux admin skills would have handled it. But that’s not me.
What is “PBO 2”?
By the way, you are correct that the Windows plotter is slower; significantly slower.
Some months ago, I was using my i5 laptop to create k34 plots.
My first attempts were via a Linux boot flash drive. I chose MX Linux.
Sometime later, I discovered that for Windows, the only file you need is “chia_plot_k34.exe” (there are no other dependencies). I was reluctant to install Chia on my Windows laptop, and now I did not have to. I needed only the one executable.
Well, it took way, way longer to complete a k34 plot, compared to MX Linux and ext4.
When I started out with Chia, I knew none of the above. So I went with Windows.
Linux desktop experience is the same/similar as Windows for at least 10 years. Monkey can install it, and run it.
PBO is that automatic overclocking crap in BIOS.
If you do not know basic stuff, how do you want to play with IT?
k34 plots is same bollock like Windows plotting…you lose 33%
That is true.
Also true is that if you asked 1,000 Windows users how to install Linux, 99% of them would probably say “What’s Linux?”.
People have to know that the installation is easy, in order to know that the installation is easy. But they don’t know.
And none of that would help casual computer users to troubleshoot the Gigabyte USB issue with the B550 motherboards.
It seems basic after you become experienced.
Playing with the gear is one way to gain experience.
I think that it is closer to 6%.
how did you figure out?
it is well known firmware issue for at least two years…affecting all 5xx series AMD motherboards. If you play with new tech, you have to follow news…especially, with weird PC behaviour…you wait for new BIOS/AGESA FW like an addict for drugs
My ASUS X570 MB got at least 10x new BIOS in last two years…AMD Ryzen 3xxx/5xxx vs PCIe 4.0 vs USB issues. Forums were full of whining
It has nothing to do with OS
I think that it is closer to 6%.
know is do - if you must think, it means you have no idea while trying to figure out solution for the problem. It is well-known fact that poor refuse to understand
Yes, the OS has nothing to do with that USB related BIOS issue.
But non tech savvy people will not know how to determine the trigger or root cause of the USB disconnecting issue.
If you were using an OS that is foreign to you, then you would have that to deal with. You would not know that the OS was not responsible. You would know only that your USB drives were disconnecting.
The above is only one example of a problem that could arise.
That is the reason for using the OS with which you have experience.
Recommending to people to use an OS that they do not know, and start crypt-o mining for the first time, is irresponsible.
[quote=“hajes29a, post:16, topic:18335”]
I think that it is closer to 6%.
How do you “know” that it k34 plots lose 33%, compared to k32 plots?
What is your source?
I have simple strategy for at least 30 years DO NOT USE WHAT YOU DON’T UNDERSTAND. If you cannot even start PC, what do you want to do with it? I recall teaching my IT teacher about IT in school in 1995 ROFL She had no idea about new tech, yet they forced her to teach it 
If you wanna play with something RTFM, UTFG…30 years ago we had just classical RTFM. Nowadays, almost everything is on Internet…yet people are dumbest in whole history, too lazy to learn anything or even use brain.
simple troubleshooting is like teaching cat barking. if you cannot do it, why the hell would you even try? I recall, most couldn’t even start/shut down PC 30 years ago or even install Windows…process is same since Win 1.0 Just nowadays, you have all basic drivers already in.
read the post, it has been run by various people. you can try it yourself. plotting anything over k32 makes no sense…once, it will be mandatory…replotting will be faster 10 years later than is now. Why the hell would you waste your resources?
I plot for 0.05$/plot…it takes 2 months just to repay plot now. Why would I want 3 months? HDD ROI is now 6-8 years. Why waste resources in low profit times?
Or are you one of those lambo kids waiting for BTC 200k $, and Chia back to rug-pull times of 1600$ ?
twenty characters…
Will just add my 2c on this thread. I started with Chia on Win like many folks and ran in to USB issues as soon as I started moving a lot of data from on-board NVMe or SSDs to external USB 3.0 HDDs especially when I was copying to 4+ disks at a time. Win logs were no help as the external disks appeared to randomly disconnect. I thought it may have been an issue with my 16-port USB hub but I changed that and issues persisted. Heck I even RMA’ed my brand new Dark Hero X570 MB and replacement was doing the same thing. I then tried Linux and started to monitor the kernel log. Kernel log in Linux started to give me hints on what may be causing the issue as the problem was reproducible in Linux as well but at least there I had more detail info. On my specific MB I finally found a fix/workaround to this issue which had something to do with IOMMU and if I recall correctly I had to disable it which resolved the problem. I hope this helps with your issue. IOMMU is an AMD specific virtualization tech so if you are running Intel CPU then it is likely not relevant. One other thing to note is that most USB controllers have a limit of how many total devices they will support. If you go over the limit the last devices connected/detected will not be reachable or other issues may occur even if there is no clearly defend limit in the USB controller firmware like random device disconnections.
On the other topic regarding Linux vs Windows, the above experience started to turn me away from Win. Once MadMax came out and I’ve noticed the performance loss of ~20% on Win I completely ditched Win and switch to Linux as my daily home driver. If you are a noob to Linux I suggest you try Linux Mint (/w Cinamon) or possibly PopOS! (Ubuntu based) both of which should be fairly easy for beginners although you cannot go wrong with pure Ubuntu as you will find out that software/hardware vendors that support Linux will likely support Ubuntu among some other popular Linux variants. If you know nothing about Linux then hop on YouTube and start searching for how-to videos as there are plenty just try to focus on videos made within the last year or so as just like with any other tech things become obsolete quickly and you may be pulling your hair out for nothing following old/outdated instructions. At this time I have a dual boot system with Linux and Win 11 on another NVMe but Win 11 is rarely booted. I only keep Win 11 for when I have to update some HW firmware and vendor releases that for Win only 
Oh and my fastest single K32 plot on Linux with 5950X & MaxMax was just over 1100sec not counting transfer to perm HDD. Lastly, in my honest opinion as an IT pro who has been using Win, Mac, Linux and some other less know OS’es for decades I suggest that everyone tries Linux as Win will only get much worse from now on, as already illustrated by Win 11 and it’s slow adoption, since MS openly stated they are not in the OS business anymore and just like Google they want to be in marketing business which means they want to sell to you based on your product usage including Win usage. There is A LOT of telemetry sent back to MS already and Win 11 starts to show where MS is headed. Win 11 search is just one example while Edge browser will keep bugging you to login with a MS account so they can keep tracking you. Just compare the license agreements between Win 7, Win 10 (arbitration clause hidden which is likely why MS offered free upgrades for a LONG time), and Win 11 (arbitration clause is the first sentence in BOLD!). Bottom line is that by staying with Win you are paying MS to be their beta tester! Many stats (ex: Steam gaming stats) show that Linux adoption is increasing, whether that is because of recent advancements which improve support for gaming on Linux or other reasons Win is slowly dying.
