Open JonathanGorr opened 3 years ago
I thought threads meant total, not per plot. Reducing it to see if that fixes it.
Reducing the threads to 2 or 4 didnt fix it.
There might be something wrong with my ram, as i'm getting bad allocation errors.
Try to increase your memory_buffer. It should not be below 3400 MB.
/Edit: What's the size of your SSD?
I have the same issue. Sometimes my plots just disappear from the jobs, without any error. It happens to at least 15% of my plots, it really sucks. I have enough RAM (32GB), 2x2TB SSD, not running out of any resource. It started happening on my 2nd rig as well, although less often.
@TegrityFarm Okay. I swapped for brand new ram and the problem still happens. I will try adding all my ram back to increase the total available. I have 2x 2 tb SSDs.
I meant to increase your Ram (memory_buffer) in your config.
I'm fighting with same issue. I have few harvester PC for making plots and only 2 PC have this problem. I tested hardware with AIDA64 stress test for 3 hours, also completed with memtes86 4/4 pass with "0" error. But i lost some plots work on this PC. Cache/temp folder have temp files for this plots also, logs for this plots stopped without error. If i don't deleted this vanished plots temp files, my Cache SSD can be full.
Look at the pic, plot-k32-2021-05-27-08-16... and plot-k32-2021-05-27-17-13... plot job vanished and this plot temp files still here:
Log from plot-k32-2021-05-27-08-16... stopped without any info or error:
Log from plot-k32-2021-05-27-17-13... stopped with this error: (First time i see this error on my vanished plot works. Others like first example, without error)
Bucket 70 uniform sort. Ram: 3.498GiB, u_sort min: 1.500GiB, qs min: 0.479GiB.
Bucket 71 uniform sort. Ram: 3.498GiB, u_sort min: 1.500GiB, qs min: 0.479GiB.
Bucket 72 uniform sort. Ram: 3.498GiB, u_sort min: 1.500GiB, qs min: 0.479GiB.
Bucket 73 uniform sort. Ram: 3.498GiB, u_sort min: 1.500GiB, qs min: 0.479GiB.
Bucket 74 uniform sort. Ram: 3.498GiB, u_sort min: 1.500GiB, qs min: 0.479GiB.
Caught plotting error: remove: The process cannot access the file because it is being used by another process.: "Z:\CACHE\plot-k32-2021-05-27-17-13-0f40b9f9cf518f4ab2a446bdce7e1035e2fa865bffd62e109a9f95bfb1a153ca.plot.p3.t6.sort_bucket_074.tmp"
[5936] Failed to execute script chia
Traceback (most recent call last):
File "chia\cmds\chia.py", line 81, in
So guys, I solved it for myself. Now, I did not have time to test each possible solution separately so I changed everything at once. My specs: Ryzen 9 5900X, 32GB 3600 MHz RAM, 2x2TB Gen 4 SSD My former, problematic settings: 12 plots limit, 30min time between plots, total max 4 phase 1, max 6 per SSD, Windows priority: real time (this is my guess that was causing the issue, also UNIX priority was problematic on Ubuntu when adjusted to maximum priority), 4000MB RAM per plot, 16 threads per plot My new settings that removed the issue (so far it seems) - I also increased amount of parallel plots but that was to increase the load on my CPU:
BTW if anyone was curious, I am getting about 35-40 plots per day with these new settings (32 for the old ones, minus about 8 disappeared plots per day, which makes it 24).
Hi,
Once I get above 6 or 7 plots, the 7th or 8th plot will enter stage 1, be worked until about 10% or less, then vanish! This is really a waste of electricity and I can't understand why it happens. I want to run 12-15 concurrently, but this makes it impossible. I am not maxxed out on threads, space or ram FYI. Here is my job: