Many thanks to MoneroOcean for his job.
Basic system is forked from last single coin repository:
Coin configs, base.sql and deployments are modified by me to use from deployment scripts
The nodejs-pool is built around a small series of core daemons that share access to a single LMDB table for tracking of shares, with MySQL being used to centralize configurations and ensure simple access from local/remote nodes. The core daemons follow:
api - Main API for the frontend to use and pull data from. Expects to be hosted at /
remoteShare - Main API for consuming shares from remote/local pools. Expects to be hosted at /leafApi
pool - Where the miners connect to.
longRunner - Database share cleanup.
payments - Handles all payments to workers.
blockManager - Unlocks blocks and distributes payments into MySQL
worker - Does regular processing of statistics and sends status e-mails for non-active miners.
API listens on port 8001, remoteShare listens on 8000
Xmrpool.net (The refrence implementation) uses the following setup:
- https://xmrpool.net is hosted on it's own server, as the main website is a static frontend
- https://api.xmrpool.net hosts api, remoteShare, longRunner, payments, blockManager, worker, as these must all be hosted with access to the same LMDB database.
Sample Caddyfile for API:
https://api.xmrpool.net {
proxy /leafApi 127.0.0.1:8000
proxy / 127.0.0.1:8001
cors
gzip
}
It is critically important that your webserver does not truncate the /leafApi
portion of the URL for the remoteShare daemon, or it will not function! Local pool servers DO use the remoteShare daemon, as this provides a buffer in case of an error with LMDB or another bug within the system, allowing shares and blocks to queue for submission as soon as the leafApi/remoteShare daemons are back up and responding with 200's.
- 4 Gb Ram
- 2 CPU Cores (with AES_NI)
- 60 Gb SSD-Backed Storage - If you're doing a multi-server install, the leaf nodes do not need this much storage. They just need enough storage to hold the blockchain for your node. The pool comes configured to use up to 24Gb of storage for LMDB. Assuming you have the longRunner worker running, it should never get near this size, but be aware that it /can/ bloat readily if things error, so be ready for this!
- Notably, this happens to be approximately the size of a 4Gb linode instance, which is where the majority of automated deployment testing happened!
- If you're planning on using e-mail, you'll want to setup an account at https://mailgun.com (It's free for 10k e-mails/month!), so you can notify miners. This also serves as the backend for password reset emails, along with other sorts of e-mails from the pool, including pool startup, pool Monerod daemon lags, etc so it's highly suggested!
- Pre-Generate the wallets, or don't, it's up to you! You'll need the addresses after the install is complete, so I'd suggest making sure you have them available. Information on suggested setups are found below.
- If you're going to be offering PPS, PLEASE make sure you load the pool wallet with XMR before you get too far along. Your pool will trigger PPS payments on it's own, and fairly readily, so you need some float in there!
- Make a non-root user, and run the installer from there!
- Add your user to
/etc/sudoers
, this must be done so the script can sudo up and do it's job. We suggest passwordless sudo. Suggested line:<USER> ALL=(ALL) NOPASSWD:ALL
. Our sample builds use:pooldaemon ALL=(ALL) NOPASSWD:ALL
- Run the deploy script as a NON-ROOT USER. This is very important! This script will install the pool to whatever user it's running under! Also. Go get a coffee, this sucker bootstraps the monero installation. Install Script: curl -L https://raw.githubusercontent.com/arqma/nodejs-pool-arqma/master/deployment/deploy_arqma.bash | bash
- Once it's complete, change as
config.json
appropriate. It is pre-loaded for a local install of everything, running on 127.0.0.1. This will work perfectly fine if you're using a single node setup. You will also want to run: source ~/.bashrc This will activate NVM and get things working for the following pm2 steps. - You'll need to change the API end point for the frontend code in the
poolui/build/globals.js
andpoolui/build/global.default.js
-- This will usually behttp(s)://<your server FQDN>/api
unless you tweak caddy! - Check
config.json
and change as appropriate. The default database directory/home/<username>/pool_db/
is already been created during startup. If you change thedb_storage_path
just make sure your user has write permissions for new path. Run:pm2 restart api
to reload the API for usage. You'll also want to setbind_ip
to the external IP of the pool server, andhostname
to the resolvable hostname for the pool server.pool_id
is mostly used for multi-server installations to provide unique identifiers in the backend. - Hop into the web interface (Should be at
http://<your server IP>/#/admin
), then login withAdministrator/Password123
, MAKE SURE TO CHANGE THIS PASSWORD ONCE YOU LOGIN. <- This step is currently not active, we're waiting for the frontend to catch up! Head down to the Manual SQL Configuration to take a look at what needs to be done by hand for now. - From the admin panel, you can configure all of your pool's settings for addresses, payment thresholds, etc.
- Once you're happy with the settings, go ahead and start all the pool daemons, commands follow.
cd ~/nodejs-pool/
pm2 start init.js --name=blockManager --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=blockManager
pm2 start init.js --name=worker --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=worker
pm2 start init.js --name=payments --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=payments
pm2 start init.js --name=remoteShare --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=remoteShare
pm2 start init.js --name=longRunner --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=longRunner
pm2 start init.js --name=pool --log-date-format="YYYY-MM-DD HH:mm Z" -- --module=pool
pm2 restart api
Install Script:
curl -L https://raw.githubusercontent.com/arqma/nodejs-pool-arqma/master/deployment/deploy_arqma.bash | bash
The installer assumes that you will be running a single-node instance and using a clean Ubuntu 16.04 server install. The following system defaults are set:
- MySQL Username: pool
- MySQL Password: 98erhfiuehw987fh23d
- MySQL Host: 127.0.0.1
- MySQL root access is only permitted as the root user, the password is in
/root/.my.cnf
- SSL Certificate is generated, self-signed, but is valid for Claymore Miners.
- The server installs and deploys Caddy as it's choice of webserver!
The following raw binaries MUST BE AVAILABLE FOR IT TO BOOTSTRAP:
- sudo
I've confirmed that the default server 16.04 installation has these requirements.
The pool comes pre-configured with values for Monero (XMR), these may need to be changed depending on the exact requirements of your coin. Other coins will likely be added down the road, and most likely will have configuration.sqls provided to overwrite the base configurations for their needs, but can be configured within the frontend as well.
The pool ALSO applies a series of patches: Fluffy Blocks, Additional Open P2P Connections, 128 Txn Bug Fix. If you don't like these, replace the auto-installed monerod fixes!
The pool is designed to have a dual-wallet design, one which is a fee wallet, one which is the live pool wallet. The fee wallet is the default target for all fees owed to the pool owner. PM2 can also manage your wallet daemon, and that is the suggested run state.
- Generate your wallets using
/usr/local/src/arqma/bin/arqma-wallet-cli
- Make sure to save your regeneration stuff!
- Start the wallet using PM2:
pm2 start /usr/local/src/arqma/bin/-wallet-rpc -- --rpc-bind-port 26969 --wallet-file <Your wallet name here> --password <Your wallet password here> --disable-rpc-login --trusted-daemon
- If you don't use PM2, then throw the wallet into a screen and have fun.
Pretty similar to the above, you may wish to dig through a few other things for sanity sake, but the installer scripts should give you a good idea of what to expect from the ground up.
Until the full frontend is released, the following SQL information needs to be updated by hand in order to bring your pool online, in module/item format. You can also edit the values in sample_config.sql, then import them into SQL directly via an update.
Critical/Must be done:
pool/address
pool/feeAddress
general/shareHost
Nice to have:
general/mailgunKey
general/mailgunURL
general/emailFrom
SQL import command: sudo mysql pool < ~/nodejs-pool/sample_config.sql (Adjust name/path as needed!)
The shareHost configuration is designed to be pointed at wherever the leafApi endpoint exists. For xmrpool.net, we use https://api.xmrpool.net/leafApi. If you're using the automated setup script, you can use: http://<your IP>/leafApi
, as Caddy will proxy it. If you're just using localhost and a local pool serv, http://127.0.0.1:8000/leafApi will do you quite nicely
Additional ports can be added as desired, samples can be found at the end of base.sql. If you're not comfortable with the MySQL command line, I highly suggest MySQL Workbench or a similar piece of software (I use datagrip!). Your root MySQL password can be found in /root/.my.cnf
Until the main frontend is done, we suggest running the following SQL line:
DELETE FROM pool.users WHERE username = 'Administrator';
This will remove the administrator user until there's an easier way to change the password. Alternatively, you can change the password to something not known by the public:
UPDATE pool.users SET email='your new password here' WHERE username='Administrator';
The email field is used as the default password field until the password is changed, at which point, it's hashed and dumped into the password field instead, and using the email field as a password is disabled.
You should take a look at the wiki for specific configuration settings in the system.
If upgrading the pool, please do a git pull to get the latest code within the pool's directory.
Once complete, please cd
into sql_sync
, then run node sql_sync.js
This will update your pool with the latest config options with any defaults that the pools may set.
This is likely due to LMDB's MDB_SIZE being hit, or due to LMDB locking up due to a reader staying open too long, possibly due to a software crash. The first step is to run:
mdb_stat -fear ~/pool_db/
This should give you output like:
Environment Info
Map address: (nil)
Map size: 51539607552
Page size: 4096
Max pages: 12582912
Number of pages used: 12582904
Last transaction ID: 74988258
Max readers: 512
Number of readers used: 24
Reader Table Status
pid thread txnid
25763 7f4f0937b740 74988258
Freelist Status
Tree depth: 3
Branch pages: 135
Leaf pages: 29917
Overflow pages: 35
Entries: 591284
Free pages: 12234698
Status of Main DB
Tree depth: 1
Branch pages: 0
Leaf pages: 1
Overflow pages: 0
Entries: 3
Status of blocks
Tree depth: 1
Branch pages: 0
Leaf pages: 1
Overflow pages: 0
Entries: 23
Status of cache
Tree depth: 3
Branch pages: 16
Leaf pages: 178
Overflow pages: 2013
Entries: 556
Status of shares
Tree depth: 2
Branch pages: 1
Leaf pages: 31
Overflow pages: 0
Entries: 4379344
The important thing to verify here is that the "Number of pages used" value is less than the "Max Pages" value, and that there are "Free pages" under "Freelist Status". If this is the case, them look at the "Reader Table Status" and look for the PID listed. Run:
ps fuax | grep <THE PID FROM ABOVE>
ex:
ps fuax | grep 25763
If the output is not blank, then one of your node processes is reading, this is fine. If there is no output given on one of them, then proceed forwards.
The second step is to run:
pm2 stop blockManager worker payments remoteShare longRunner api
pm2 start blockManager worker payments remoteShare longRunner api
This will restart all of your related daemons, and will clear any open reader connections, allowing LMDB to get back to a normal state.
If on the other hand, you have no "Free pages" and your Pages used is equal to the Max Pages, then you've run out of disk space for LMDB. You need to verify the cleaner is working. For reference, 4.3 million shares are stored within approximately 2-3 Gb of space, so if you're vastly exceeding this, then your cleaner (longRunner) is likely broken.
If you're considering PPS, I've spoken with Fireice_UK whom kindly did some math about what you're looking at in terms of requiements to run a PPS pool without it self-impoloding under particular risk factors, based on the work found here
Also I calculated the amount of XMR needed to for a PPS pool to stay afloat. Perhaps you should put them up in the README to stop some spectacular clusterfucks :D:
For 1 in 1000000 chance that the pool will go bankrupt: 5% fee -> 1200 2% fee -> 3000
For 1 in 1000000000 chance: 5% fee -> 1800 2% fee -> 4500
The developers of the pool have not verified this, but based on our own usage on https://xmrpool.net/ this seems rather reasonable. You should be wary if you're consdering PPS and take you fees into account appropriately!
If you need help installing the pool from scratch, please have your servers ready, which would be Ubuntu 16.04 servers, blank and clean, DNS records pointed. These need to be x86_64 boxes with AES-NI Available.
SSH access with a sudo-enabled user will be needed, preferably the user that is slated to run the pool.
Assistance is not available for frontend customization at this time.
For assitance, please contact ArqTras at [email protected]
If you'd like to make a one time donation, the addresses are as follows:
* AEON WmtWzE4zej7FLjctgMPk7Va7hiiESJF5xHfmb5KZaMAhHDSRU51pqTJQiVFZgRMmrga9KvqPg48EZTEVQ9qA47o52MYgiMqaw
* XMR 4AYuDc4cEqxfxVTUFwVqPd4JdmKLjm9dNhTjuT6Ud5gQa564wp6cxMBWbwaVe4vUMveKAzAiA4j8xgUi29TpKXpm3zc3jmn
* ETN etnkLgWfr5uE8MZSZpsabb6HjG8Mig9qaS4wQ6Hu2VVKTiJT9Ucdrzz9CqGF9tycaWbntrSRr1CwVJDqGYPtumL72GhXwtoPvs
- MoneroOcean - Based pool system with many important modifications
- Zone117x - Original node-cryptonote-pool from which, the stratum implementation has been borrowed.
- Miziel - Frontend mod XMRPoolUI
- Wolf0/OhGodAGirl - Rebuild of node-multi-hashing with AES-NI * * node-multi-hashing
- [Snipa22] https://github.com/Snipa22/nodejs-pool This awsome pool software