guy.davis
-
Posts
334 -
Joined
-
Last visited
-
Days Won
1
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by guy.davis
-
-
Machinaris v2.2.0 is now available:
Added
- Support for Achi blockchain. Requires its own plots (like Chives), you'll have to plot separately if interested. Thanks @priyankub
- Optionally launch Chia Data Layer services if env var `chia_data=true` is set.
Changed
- Harvester mode will now optionally also run chia-exporter for Prometheus results.
Updated
- Chia to v2.2.0 - misc improvements, see their release notes.
- Cactus to v2.1.4.
- Gigahorse to v2.1.4.giga26.
- Wheat to v2.1.6
Notes
- Support for new blockchains and tools DOES NOT imply my endorsement for them. Only run those you are comfortable with.
UPDATE: CNI corporate retracted their v2.2.0 release due to nasty bugs in their database code. As such, I had to pull Machinaris v2.2.0 until they get their issues sorted out. -
Machinaris version 2.1.4 is now available.
Added
- Optionally launch `chia-exporter` [metrics endpoint](https://github.com/Chia-Network/chia-exporter) for Prometheus reporting if env var `chia_exporter=true` is set.
- Support for running Gigahorse containers on arm64 architectures such as Raspberry Pi.
- Main dashboard page can now be pinned (Pin Icon) to display only a blockchain of interest, rather than rotating through each in turn.
Changed
- Optionally launch of Gigahorse recompute server when in harvester mode if env var `gigahorse_recompute_server=true` is set.
- Fix for Bladebit GPU plotting jobs taking a minute to display on Plotting page on job start. Should now be seconds only.
Updated- [Chia](https://github.com/Chia-Network/chia-blockchain/releases/tag/2.1.4) to v2.1.4 - misc improvements, see their release notes.
- [Gigahorse](https://github.com/madMAx43v3r/chia-gigahorse/releases/tag/v2.1.3.giga26) to v2.1.3.giga26.
-
On 1/8/2024 at 11:34 AM, Einsteinjr said:
FYI - for a while now, I've been getting much higher skipped SPs than normal (>20%). Did a bit of research and debugging the logs and found that modifying [machinaris.Config]/mainnet/config/config.yaml file fixed the issue.
target_peer_count needs to be changed from 80 to 20 and the errors I was seeing seemed to have disappeared and the skipped rates is <1%.
Good point. Yes, I run at 10 to minimize memory usage for synced blockchains. Details on the wiki.
-
Version 2.1.3 now available that addresses the unfortunate bug CNI released in v2.1.2. Details at: https://github.com/guydavis/machinaris/releases/tag/2.1.3
-
Version 2.1.2 now available.
### Added
- Settings | Plotting | Bladebit now supports "no_direct_io: true" option
- Settings | Plotting | Bladebit now supports "check_plots: 100" option to check at end of plot creation
- Settings | Plotting | Bladebit now supports "check_threshold: 0.6" option to delete invalid plots at end of plot creation
### Changed
- Support for Gigahorse Recompute Server (single GPU on fullnode/farmer), shared by multiple Gigahorse harvesters.
### Updated
- Chia to v2.1.2 - misc improvements, see their release notes.
- Gigahorse to v2.1.1.giga25 with C20 compression support.
- Flask to v3.0.0 - serves up the Machinaris WebUI.
### Notes
- Regression in Chia binaries (since v2.0.0) ignoring the "-y" parameter on plotnft changes, has broken pool join/leave thru WebUI.
-
On 11/23/2023 at 9:44 AM, DoeBoye said:
I'm having problems deleting a cold wallet address I mistakenly added to the Chinilla fork. On the wallet page, I open the edit wallet option, delete the key for the cold wallet, save and close. When I reopen the edit option, the wallet key is still there.
Am I doing something wrong, or is this a bug? Is there somewhere I can manually remove it?
Go into the shell of the `machinaris` Chia fullnode container and delete file:
/root/.chia/machinaris/config/cold_wallet_addresses.jsonThen add your address back, just for the Chia row, using the Machinaris WebUI's wallets page.
-
1
-
-
On 11/20/2023 at 5:32 AM, radiumx3 said:
@gerard4 did you found out the issue ? Can't figure out what is wrong..
Due to a regression in the Chia binaries provided by CNI, they no longer honor the "-y" flag allowing the commandline to be run by the WebUI. You'll need to run the "chia plotnft create" command (see log for exact parameters) yourself within the shell of the container as you will be prompted for a manual confirmation. Sorry for the trouble this defect from the CNI group has caused.
-
On 10/28/2023 at 10:32 AM, 666sniperererer said:
Hello, I am a bit hopeless now and really need to ask for help...
I got Tesla P4 in my Unraid tower, all settings to Machinaris container seem to be OK (I am able to see the card via "nvidia-smi" inside the container), but it is not being used at all - Machinaris tries to decompress (C7) using only CPU.
I run the same absolutely same settings as on my previous server (with the same Tesla P4):
harvester:
chia_ssl_ca:
crt: config/ssl/ca/chia_ca.crt
key: config/ssl/ca/chia_ca.key
decompressor_thread_count: 1
decompressor_timeout: 20
disable_cpu_affinity: false
enforce_gpu_index: false
farmer_peer:
host: 127.0.0.1
port: 8447
gpu_index: 0
logging: *id001
max_compression_level_allowed: 7
network_overrides: *id002
num_threads: 30
parallel_decompressor_count: 1
parallel_read: true
plot_directories:
- /plots1
- /plots2
plots_refresh_parameter:
batch_size: 300
batch_sleep_milliseconds: 1
interval_seconds: 120
retry_invalid_seconds: 1200
port: 8448
private_ssl_ca:
crt: config/ssl/ca/private_ca.crt
key: config/ssl/ca/private_ca.key
recursive_plot_scan: false
rpc_port: 8560
selected_network: mainnet
ssl:
private_crt: config/ssl/harvester/private_harvester.crt
private_key: config/ssl/harvester/private_harvester.key
start_rpc_server: true
use_gpu_harvesting: true
inbound_rate_limit_percent: 100And in log this repeats:
2023-10-28T18:34:37.200 harvester chia.harvester.harvester: WARNING No decompressor available. Cancelling qualities retrieving for /plots2/plot-k32-c07-2023-10-14-00-57-ef88c41d8f61deda9d44059d58a4085a764392813d497cc24dd4f65f797470f5.plot
I will be very grateful for advice, thanks in advance!
All I can suggest is a reset of the Chial blockchain config to get back to a default state. This assumes you have already installed the Nvidia plugin for Unraid.
-
On 10/20/2023 at 8:06 PM, Shotkar said:
Hi,
I'm having trouble with pooling I seem to have two pools in pools and am self pooling in settings|pools and get an error when changing. I have reinstalled Machinaris and also Uninstalled Machinaris and then manually deleted all files I could find. Also have used the command console "chia plotnft leave". But these pools seem to keep showing up with each new install. Is there some file that i am missing to fully reset?
Thanks
Yes, there is an unfortunate regression in the Chia command-line binaries that breaks automation. CNI ("Chia Network Inc." may fix it one day). Workaround is to issue the `chia plotnft ... ` commands from within the container's shell.
-
7 minutes ago, mishmash- said:
Hello all, I've searched the docs and the config files but cannot find it - how can we remove old forks from the machinaris list? I have an old btcgreen fork in my list I'd like to remove...
Good question. I have now documented the removal of old/stale/unwanted workers.
-
Machinaris v2.1.0 now available: https://github.com/guydavis/machinaris/releases/tag/2.1.0
-
Bladebit to v3.1.0 - Supporting hybrid GPU/disk plotting with either 128 GB or 16 GB RAM.
-
Chia to v2.1.0 - misc improvements, see their release notes.
-
-
On 8/25/2023 at 8:33 AM, jhewit485 said:
Can you advise why c3 compressed plots would not be appearing in the count of plots? Generated them using plotman, with bladebit ramplot. Thanks in advance. They are on the hard drive but have c3 in the plots title but seem to be ignored by scraper and also the plot checker.
Reading CNI release notes from their betas, it seems they recommend generate a brand-new config.yaml to add all their new settings. So try this procedure to reset your Chia config.
-
Machinaris v2.0.0 is now available.
-
14 hours ago, Normand_Nadon said:
hello @guy.davis and thank you for your work.
I was wondering, why is Machinaris taking so much ressources on my Unraid Server?I am farming only, and the hard drives are in the same machine, mounted directly inside the container.
Fuse is not used for those drives, they are directly mounted as /mnt/disks/farmer## (4 x 6GB HDD)Machinaris takes anywhere from 6 to 8 GB of RAM and makes regular 100% spikes on cpu cores...
What explains this behavior?Please see the wiki to scale down. I follow most of these to run on my older hardware.
-
On 7/9/2023 at 1:29 PM, tutulino said:
Hello, how can I update to chia 1.8.2 client? pool is rejecting my shares.
At unraid app is shown as updated, but chia inside is 1.7.0
Hi, please try a 'Force Update' from the Docker tab of the Unraid Admin UI. Also, ensure you are running a recent, supported version of Unraid. Details here.
-
Machinaris v1.8.2 now available: https://github.com/guydavis/machinaris/releases/tag/1.8.2
-
1
-
-
Machinaris v1.8.1 now available.
Changed
- By default, exclude Gigahorse plots from automated plot check due to poor performance of the 'ProofOfSpace' binary.
- Fix for plotting manager issues when plotting with multiple tmp drives.
Updated
- Bladebit to 3.0.0-alpha4. See known issues.
- BPX to v2.1.0
- Cactus to v1.8.0
- Chia to v1.8.1
- Flora to v1.8.0, updated by ageorge95.
- Gigahorse to v1.8.0.giga11.
Notes
- Bladebit still broken on ability to ramplot and diskplot as they try to match Gigahorse's GPU plotting. Still awaiting a fix...
- Support for new blockchains and tools DOES NOT imply my endorsement for them. Only run those you are comfortable with.-
1
-
-
18 hours ago, Shotkar said:
Hi, just wondering how I go about setting the Optional config setting to restart fork fullnodes if they consume too much memory?
Have looked though all the configs but cant seem to find it. Thanks.
Hi, please point #3 on this wiki page. However, I will say it's not a panacea, as I've found some fullnodes from very out-of-date chains like Flora, N-Chain, and Stor simply can not stay under 3 GB for long. For some, I found they were simply restarting over and over unless I gave them 4+ GB. So, this setting can help, but for the worst offenders, you may just want to skip farming that particular chain.
-
1
-
-
Good day, the next version of Machinaris (v1.8.0) is now available:
Added
- Concurrent plot transfers to both remote and local archive paths. Previously was one plot transferred at a time.
- Checking for adequate free space on both tmp and dst paths before starting another plotting job.
- Replotting selection (Farming page settings) can now target uncompressed plots for deletion.
Changed
- Fix for Download button on Settings pages that were providing the wrong config file. Thanks @TuftyBruno.
- Updated German and Portugese translations. Thanks to @slowfinger and @antcasq. All contributions are appreciated.
- Follow Chia version numbers (like 1.8.0) from now on, as Machinaris is on their release cadence.
Updated
- Cactus to v1.7.1
- Chia to v1.8.0
- Gigahorse to v1.8.0.giga10.
- HDDCoin to 3.0.0-beta1 based on Chia 1.7.0.
- MMX to v0.10.2 on `testnet10`.
Notes
- Due to critical security patches in the bundled Chia v1.8.0 release, users are encouraged to upgrade as soon as possible.
- Yes, Bladebit 3 is still alpha and not reliable for plotting.-
1
-
-
On 4/27/2023 at 9:36 AM, veri745 said:
Anyone know why my Tdarr and Tdarr_node containers would be multiple gigs each?
Name Container Writable Log tdarr 5.58 GB 1.50 GB 11.7 MB tdarr_node 3.73 GB 531 MB 11.5 MBThis is from docker -> "Container Size"
I got a warning after installing and using Tdarr for a bit that my docker vdisk was almost full
*edit* I verified that my transcode cache for all of my libraries is properly mapped to a `transcode_cache` share on my cache drive, not in the image
Hi, did you ever find a solution to tdarr_node doing all its work at path inside container: /app/tdarr_node? It's completely ignoring the mapping to I set to "/temp" on a spare SSD. So, if I let it run for a little while it completely fills the Unraid Docker image. Thanks in advance.
-
Good day, the next version of Machinaris is now available. Thanks to all those that contributed fixes, testing, and other feedback.
Added:
- Support for running a blockchain timelord with environment variable `mode` set to `fullnode,timelord`.
- Optional config setting to restart fork fullnodes if they consume too much memory. Looking at you Flora, HDDCoin, N-Chain, etc...
Changed:
- On fresh install, optionally download (via libtorrent) a recent Chia database checkpoint. https://www.chia.net/downloads
- Fixes for Gigahorse GPU plotting including 64 GB RAM mode using the `tmp3` SSD plotting path.
- Fix for Gigahorse Alerts, please reset earlier broken settings file to https://github.com/guydavis/machinaris/blob/develop/config/chiadog/gigahorse.sample.yaml.
- Fix for container stop signal to cleanly shutdown forks too, as was already happening for Chia.
Updated:
- Bladebit to 3.0.0-alpha3, so still unstable. https://downloads.chia.net/bladebit/
- Chia to v1.7.1 https://github.com/Chia-Network/chia-blockchain/releases/tag/1.7.1
- Gigahorse to v1.6.2.giga7 with OpenCL support for AMD GPUs. https://github.com/madMAx43v3r/chia-gigahorse/releases/tag/v1.6.2.giga7
- MMX to v0.9.14 on `testnet9`. Supporting the latest AMD GPU driver. https://github.com/madMAx43v3r/mmx-node/releases/tag/v0.9.14
- One to v1.7.0 https://github.com/xone-network/one-blockchain/releases/tag/1.7.0
Notes:
- Bladebit has regressed on the ability to `ramplot` as they try to match Gigahorse's GPU plotting. Still awaiting a fix...
- Support for new blockchains and tools DOES NOT imply my endorsement for them. Only run those you are comfortable with.
-
Hi all! The newest version of Machinaris is now available. Thanks to everyone who provided feedback on version 0.8.7.
0.8.7 - 2023-02-16
Added- Schedule plotting on the "Settings | Plotting" page to take advantage of off-peak electricity, lower your fan noise, avoid backups, etc...
- Bladebit - alpha GPU plotting support in main Machinaris image. Set `mode: gpuplot` under `bladebit` on Settings | Plotting page.
- Gigahorse - beta GPU plotting and farming support in new Machinaris-Gigahorse image. Set `mode: gpuplot` under `madmax` on Settings | Plotting page.
- Ballcoin - blockchain at version 1.6.0, another slow Silicoin fork.
- Pipscoin blockchain at version 1.7.0.
Changed
- Fix missing Connections listing for Flax and MMX blockchains. Thanks @ekersey!
- Fix for Bladebit ramplot relaunching. Thanks @JoeZotacExperience!
- Multiple functionality & performance cleanups from excellent code review by @qoole.
- Display compression level for active plotting jobs on Gigahorse and MMX.
Updated
- BTCGreen to v1.7.0b
- Cactus to v1.6.2
- Chia to v1.7.0
- Chives to v1.5.4, including staking.
- SHIBGreen to v1.7.0.1
- Staicoin to v1.3.0. Note, they require a fresh `config.yaml`.
Notes
- Support for new blockchains and tools DOES NOT imply my endorsement for them. Only run those you are comfortable with.
-
14 hours ago, утуьнищн said:
Tell me pls, what do you think after the release of madmax gpu plotter is it possible to integrate it into the unraid system?
I want to build a system for the future, so this point is important for me, I understand that this will take time, I just want to understand purely technically is it possible to implement on an unraid system? I just don’t know much about this system, and how it works with a GPU
Good question! Yes, GPU plotting will likely be possible using the latest MMX binaries as the "machinaris-mmx" Docker image has supported GPU farming for months now. I have added initial plotting support already in the development stream, but lack an Nvidia GPU to test it myself on Unraid. MMX plot compression is working too. For the latest updates, be sure to join the Machinaris Discord.
EDIT: Above was about MMX blockchain support. Similar GPU plotting and plot compression is coming to Chia blockchain too, but has not been released publicly yet to my knowledge. Machinaris will support this on Chia with bladebit/madmax too if available.
-
1
-
-
2 hours ago, утуьнищн said:
Ok
Thanks for your patience in providing the details. For the field "Worker Address" which asks for an IP Address, you have provided a full URL including http protocol. Please use the IP address only "192.168.1.136" and restart the container.
Edit: I note you placed everything on a disk at "/mnt/disk3" including the appdata folder, plots folder, and plotting temp folder. I guess this is only for a test, on a single SSD? Please don't do this for a real farm. Appdata and plotting temp should be on separate SSDs (if you are using plotting temp, not RAM-only). I would put plots on separate HDD disks (or similar).
[Support] Machinaris - Chia cryptocurrency farming + Plotman plotting + Unraid WebUI
in Docker Containers
Posted
Version 2.2.1 of Machinaris is now available, for those that like to try the very latest Chia release from CNI corporate.
Updated
- Chia to v2.2.1 - possible fixes for serious problems in v2.2.0?
Notes
- Updated potentially buggy Chia software DOES NOT imply my endorsement. *Only upgrade when you are comfortable.*