Hi everybody. I have a system with specs below and recently started plotting Chia.
CPU: INTEL I9 10900k
M.B: ASUS Z590-F GAMING WIFI
RAM: GSKILL 64 GB BUS 4000 TRIDENT
SSD: 3* (M2 980PRO SAMSUNG 1TB)
I wanted to know if it’s worth it to stripe the three SSDs? I’m currently plotting about 21 plots a day with three job configurations below:
Not all raid controllers (hardware and software) correctly pass the TRIM commands to the drives.
A Raid 0 will just make sure every operation is held up by the slowest in the array. For SSD drives the difference can be huge because of extremes in write amplification and garbage collection.
Thank you for your answers. I tried the striping but not that much change happened. My biggest problem now is that the system freezes frequently. I have to completely turn off/on the system. Any similar experiences?
I decided to run the plotting with Chia app in three plots for each NVMe 2.0 and see if the freezing happens again.
Freezing issues can very well be caused by driver issues.
Be sure to update bios and chipset drivers.
That last bit seems to have finally solved my bsod issues - knock on wood
From the configs i see, i think you are having too many plots in phase 1 at the same time. Really want to keep that as low as possible without having plots wait for eachother
I updated the BIOS with the built-in ASUS tool. Any solution for completely checking the drivers? By the way, I don’t see BSOD. The system just freezes and no mouse/keyboard/Ctrl+Alt+Delete works. I have to turn it off from power.
Anyway, I changed the jobs and lowered phase 1 as below. I started three jobs with the code bellow and no freezes after 8 hours yet:
one of my machines is a 10850 with 2x 970 EVOs and it does that.
i run ubuntu 20.04 LTS server, I use LVM to do my software raids. Others recommend mdadm as it has slightly higher throughput, but i don’t think thats the limiting factor for chia.
so my setup is all command line and i use Plotman, but SWAR is a fine manager
I ran the swar with the config above. The freezing problem is solved, but the system restarts after two hours of working occasionally. Any solutions? I updates the chipsets as well
If event logs aren’t giving any hints it sounds like you need to go back to fundamental troubleshooting your system. Reseat all your connections. Run a full memory test. Run a 24 hour cpu stress test. consider if you might have a bad power supply and or motherboard. There are too many factors that can cause random system instability that you just need to start checking off.
As for raid 0 temp SSDs everything I’ve seen says yes, it’s all upsides with no down. It isn’t 2005, motherboards know how to handle SSD’s properly now.
I’ve updated everything, even the firmware of my SSD drives. Do you think internet connection issues may result in freezes for swar and as a result, the whole system?
I would look at non chia related issues, like suggested, run memtest64 or some other memory test, cpu stresstest, with no chia programs running and see what you get.
If it still fails then it really might be a hardware issue, such as bad Mainboard, PSU or …troubleshooting sucks. Have you tried the windows crashreports?
You know, I started some memory and disk checks. but can’t find reliable software for CPU tests. and is there any useful tool to test the health of the motherboard?
This system should easily be able to do 10 in parallel with the default chia plotter…
So assuming your final drive is an HDD and your temp is 3x1tb(raid0). Try this… it should start a plot every 30 min… which should be enough time for the plot to write to the HDD without being caught by the next plot… unless your HDD is only writing at 40MB/s or something… this would assume you are able to write at 100-200MB/s to the HDD.