[BUG] Harvester plots_refresh_parameter not keeping track of plots
See original GitHub issueThe harvester is losing track of my plots during replotting, and my plot count is going down.
I have 5156 plots. All of my drives are full, so I am currently replotting.
I have a script that, upon completion of a new NFT plot, it identifies an OG plot, deletes it, and moves the new NFT plot in its place.
I am currently creating 64 plots every 24h.
Earlier today, my harvester did indeed show that it was farming 5156
plots. I check my status a few moments ago, and it shows that I’m only farming 5129
plots.
(Then some time passes for the farmer to rediscover all my plots)
Ok, so back to 5156
after restarting the farmer. It lost track of 27 plots.
When I first discovered this bug this morning, there was a Δ of 85 plots that the harvester wasn’t tracking.
2021-08-28T13:54:20.544 harvester chia.harvester.harvester: INFO … Total 5071 plots
chia stop all -d
chia start farmer
2021-08-28T14:49:27.451 harvester chia.harvester.harvester: INFO … Total 5156 plots
What I suspect:
My config.yaml had been with me since before launch, and I decided to compare it against https://github.com/Chia-Network/chia-blockchain/blob/main/chia/util/initial-config.yaml to see if I needed to refresh it any. There were a few things that I was missing.
I had plot_loading_frequency_seconds: 120
, which isn’t anywhere in the current initial-config.yaml, so I swapped that out with: https://github.com/Chia-Network/chia-blockchain/blob/main/chia/util/initial-config.yaml#L128-L131
I think that this may not be discovering new plots for some reason. I’m going to attempt to remove those lines, and go back to plot_loading_frequency_seconds: 120
. I’ll check back in the morning to see if the count is off.
Chia: 1.2.5 Distro: Arch Linux [x86_64] Kernel: Linux 5.13.12-arch1-1 Processor: 24 x Intel® Xeon® CPU E5-2697 v2 @ 2.70GHz Processor: 24 x Intel® Xeon® CPU E5-2697 v2 @ 2.70GHz RAM: 64362 MiB
Issue Analytics
- State:
- Created 2 years ago
- Comments:11 (4 by maintainers)
Top GitHub Comments
On Linux, use rsync to copy plots if you don’t want to script changing file names. rsync creates a temporary file while transferring and renames it to the final file name after transfer is complete.
restarting the harvester alone appears to be sufficient.
chia start harvester -r