If I don't set a `electrum-rpc-addr` flag, by default electrs only listens for tcp6 connections. In order to accept tcp4, I need to add `--electrum-rpc-addr="127.0.0.1:50001`. This is not a big deal since setting the flag fixes it, but adding this info to the docs should make it easier for users. Also, the `config.rs` file should be changed since the help for this option lists `127.0.0.1:50001` as the default, which is not correct (at least on my install). I also added a sample unit file for systemd.
5.9 KiB
Installation
Install latest Rust (1.28+), latest Bitcoin Core (0.16+) and latest Electrum wallet (3.2+).
Also, install the following packages (on Debian):
$ sudo apt update
$ sudo apt install clang cmake # for building 'rust-rocksdb'
Build
First build should take ~20 minutes:
$ git clone https://github.com/romanz/electrs
$ cd electrs
$ cargo build --release
Bitcoind configuration
Allow Bitcoin daemon to sync before starting Electrum server:
$ bitcoind -server=1 -txindex=0 -prune=0
If you are using -rpcuser=USER
and -rpcpassword=PASSWORD
for authentication, please use --cookie="USER:PASSWORD"
command-line flag.
Otherwise, ~/.bitcoin/.cookie
will be read, allowing this server to use bitcoind JSONRPC interface.
Usage
First index sync should take ~1.5 hours:
$ cargo run --release -- -vvv --timestamp --db-dir ./db --electrum-rpc-addr="127.0.0.1:50001" [--cookie="USER:PASSWORD"]
2018-08-17T18:27:42 - INFO - NetworkInfo { version: 179900, subversion: "/Satoshi:0.17.99/" }
2018-08-17T18:27:42 - INFO - BlockchainInfo { chain: "main", blocks: 537204, headers: 537204, bestblockhash: "0000000000000000002956768ca9421a8ddf4e53b1d81e429bd0125a383e3636", pruned: false, initialblockdownload: false }
2018-08-17T18:27:42 - DEBUG - opening DB at "./db/mainnet"
2018-08-17T18:27:42 - DEBUG - full compaction marker: None
2018-08-17T18:27:42 - INFO - listing block files at "/home/user/.bitcoin/blocks/blk*.dat"
2018-08-17T18:27:42 - INFO - indexing 1348 blk*.dat files
2018-08-17T18:27:42 - DEBUG - found 0 indexed blocks
2018-08-17T18:27:55 - DEBUG - applying 537205 new headers from height 0
2018-08-17T19:31:01 - DEBUG - no more blocks to index
2018-08-17T19:31:03 - DEBUG - no more blocks to index
2018-08-17T19:31:03 - DEBUG - last indexed block: best=0000000000000000002956768ca9421a8ddf4e53b1d81e429bd0125a383e3636 height=537204 @ 2018-08-17T15:24:02Z
2018-08-17T19:31:05 - DEBUG - opening DB at "./db/mainnet"
2018-08-17T19:31:06 - INFO - starting full compaction
2018-08-17T19:58:19 - INFO - finished full compaction
2018-08-17T19:58:19 - INFO - enabling auto-compactions
2018-08-17T19:58:19 - DEBUG - opening DB at "./db/mainnet"
2018-08-17T19:58:26 - DEBUG - applying 537205 new headers from height 0
2018-08-17T19:58:27 - DEBUG - downloading new block headers (537205 already indexed) from 000000000000000000150d26fcc38b8c3b71ae074028d1d50949ef5aa429da00
2018-08-17T19:58:27 - INFO - best=000000000000000000150d26fcc38b8c3b71ae074028d1d50949ef5aa429da00 height=537218 @ 2018-08-17T16:57:50Z (14 left to index)
2018-08-17T19:58:28 - DEBUG - applying 14 new headers from height 537205
2018-08-17T19:58:29 - INFO - RPC server running on 127.0.0.1:50001
If initial sync fails due to memory allocation of xxxxxxxx bytes failedAborted
errors, as may happen on devices with limited RAM, try the following arguments when starting electrs
. It should take roughly 18 hours to sync and compact the index on an Odroid HC1 with 8 cpu cores @ 2GHz, 2GB RAM, and an SSD using the following command.
$ cargo run --release -- -vvvv --index-batch-size=10 --jsonrpc-import --db-dir ./db --electrum-rpc-addr="127.0.0.1:50001" [--cookie="USER:PASSWORD"]
The index database is stored here:
$ du db/
38G db/mainnet/
Electrum client
# Connect only to the local server, for better privacy
$ ./scripts/local-electrum.bash
+ ADDR=127.0.0.1
+ PORT=50001
+ PROTOCOL=t
+ electrum --oneserver --server=127.0.0.1:50001:t
<snip>
You can persist Electrum configuration (see ~/.electrum/config
) using:
$ electrum setconfig oneserver true
$ electrum setconfig server 127.0.0.1:50001:t
$ electrum # will connect only to the local server
SSL connection
In order to use a secure connection, you can also use NGINX as an SSL endpoint by placing the following block in nginx.conf
.
stream {
upstream electrs {
server 127.0.0.1:50001;
}
server {
listen 50002 ssl;
proxy_pass electrs;
ssl_certificate /path/to/example.crt;
ssl_certificate_key /path/to/example.key;
ssl_session_cache shared:SSL:1m;
ssl_session_timeout 4h;
ssl_protocols TLSv1 TLSv1.1 TLSv1.2 TLSv1.3;
ssl_prefer_server_ciphers on;
}
}
$ sudo systemctl restart nginx
$ electrum --oneserver --server=example:50002:s
Sample Systemd Unit File
You may wish to have systemd manage electrs so that it's "always on." Here is a sample unit file (which assumes that the bitcoind unit file is bitcoind.service
):
[Unit]
Description=Electrs
After=bitcoind.service
[Service]
WorkingDirectory=/home/bitcoin/electrs
ExecStart=/home/bitcoin/electrs/target/release/electrs --db-dir ./db --electrum-rpc-addr="127.0.0.1:50001"
User=bitcoin
Group=bitcoin
Type=simple
KillMode=process
TimeoutSec=60
Restart=always
RestartSec=60
[Install]
WantedBy=multi-user.target
Docker
$ docker build -t electrs-app .
$ docker run --network host \
--volume $HOME/.bitcoin:/home/user/.bitcoin:ro \
--volume $PWD:/home/user \
--rm -i -t electrs-app \
electrs -vvvv --timestamp --db-dir /home/user/db
Monitoring
Indexing and serving metrics are exported via Prometheus:
$ sudo apt install prometheus
$ echo "
scrape_configs:
- job_name: electrs
static_configs:
- targets: ['localhost:4224']
" | sudo tee -a /etc/prometheus/prometheus.yml
$ sudo systemctl restart prometheus
$ firefox 'http://localhost:9090/graph?g0.range_input=1h&g0.expr=index_height&g0.tab=0'