electrs on RaspiBlitz - running on the RaspberryPi 3

[Himbeergeld]

Hello @openoms, just tried to install electrs on my RaspiBlitz V1.1 using your "electrs_install_on_RaspiBlitz.sh" script but I get stuck due to Connection refused errors:

---

Running target/release/electrs -vvvv --index-batch-size=10 --jsonrpc-import --db-dir /mnt/hdd/electrs/db '--cookie=raspibolt: <myPsssowordhere>' '--electrum-rpc-addr=0.0.0.0:50001'
Config { log: StdErrLog { verbosity: Trace, quiet: false, timestamp: Off, modules: [], writer: "stderr", color_choice: Auto }, network_type: bitcoin, db_path: "/mnt/hdd/electrs/db/mainnet", daemon_dir: "/home/admin/.bitcoin", daemon_rpc_addr: V6([::1]:8332), cookie: Some("raspibolt:\u{1b}[C\u{1b}[C\u{1b}[C\u{1b}[D"), electrum_rpc_addr: V4(0.0.0.0:50001), monitoring_addr: V6([::1]:4224), jsonrpc_import: true, index_batch_size: 10, bulk_index_threads: 4, tx_cache_size: 10000, txid_limit: 100, server_banner: "Welcome to electrs (Electrum Rust Server)!" }
DEBUG - Server listening on [::1]:4224
DEBUG - Running accept thread
WARN - failed to connect daemon at [::1]:8332: Connection refused (os error 111)
WARN - failed to connect daemon at [::1]:8332: Connection refused (os error 111)
.... keeps printing the Connection refused messages

---

• does the script need additional "ufw allow" commands?
• is it really optimized for the RaspiBlitz (just wondering because of the "--cookie=raspibolt:.." statement above
  Thanks for your help!

[openoms]

Hi, thanks for contacting.
At which stage you are in the setup?

I ran your command and it works on my node.

It seems that electrs is not able to connect to bitcoind.
Is it running? Check with sudo systemctl status bitcoind

Do you have server=1 in you bitcoin.conf?
Run sudo nano /mnt/hdd/bitcoin/bitcoin.conf to check.
That is where you find your RPC user (raspibolt by default, because that is what the RaspiBlitz is based on. And also can confirm your PASSWORD_B.

You don`t need to allow anything else than the port 50001 through ufw, because the rest of the communication is happening inside the node.

[Himbeergeld]

Yes, "bitcoind.service" reports active (running) since more than one week. The # Bitcoind options section has "server=1" set. The problem happens at the last step of the script:
cargo run --release -- -vvvv --index-batch-size=10 --jsonrpc-import --db-dir /mnt/hdd/electrs/db --cookie="raspibolt:$PASSWORD_B" --electrum-rpc-addr="0.0.0.0:50001"
I'd expect it to be the longest part of the script but the Connection refused error appears already shortly after this command.

[openoms]

I can`t reproduce the error unfortunately.

Running target/release/electrs -vvvv --index-batch-size=10 --jsonrpc-import --db-dir /mnt/hdd/electrs/db '--cookie=raspibolt: myPsssowordhere' '--electrum-rpc-addr=0.0.0.0:50001'

in this one there a couple unnecessary `-s and the raspibolt: part needs to be between " ".

Try running just these two lines in the ~/electrs directory (type your PASSWORD_B followed by [ENTER]) :

read PASSWORD_B
cargo run --release -- -vvvv --index-batch-size=10 --jsonrpc-import --db-dir /mnt/hdd/electrs/db --cookie="raspibolt:$PASSWORD_B" --electrum-rpc-addr="0.0.0.0:50001"

to make sure the format is correct.

I get:

Config { log: StdErrLog { verbosity: Trace, quiet: false, timestamp: Off, modules: [], writer: "stderr", color_choice: Auto }, network_type: bitcoin, db_path: "/mnt/hdd/electrs/db/mainnet", daemon_dir: "/home/admin/.bitcoin", daemon_rpc_addr: V4(127.0.0.1:8332), cookie: Some("raspibolt:PASSWORD_B"), electrum_rpc_addr: V4(0.0.0.0:50001), monitoring_addr: V4(127.0.0.1:4224), jsonrpc_import: true, index_batch_size: 10, bulk_index_threads: 8, tx_cache_size: 10000, txid_limit: 100, server_banner: "electrs run by openoms" }
DEBUG - Server listening on 127.0.0.1:4224
DEBUG - Running accept thread
INFO - NetworkInfo { version: 170001, subversion: "/Satoshi:0.17.0.1/" }
INFO - BlockchainInfo { chain: "main", blocks: 569456, headers: 569456, bestblockhash: "0000000000000000001a32d96169573c7e4a8a5ff3a5047b223d2a34c3fa89db", pruned: false, initialblockdownload: false

You don`t have the default localhost (127.0.0.1) associated for some reason.

Can try adding these options to your command:
--daemon-rpc-addr="127.0.0.1:8332" --monitoring-addr="127.0.0.1:4224"

[Himbeergeld]

Thank you, the additional localhost options allowed me to proceed. I also had to modify the path variable with PATH=$PATH:$HOME/.cargo/bin:$HOME/electrs before I ran the cargo command from the ~/electrs directory. Now it seems to proceed well and is currently indexing the database, I can see the Free Mem decreasing periodically down to red level on the RaspiBlitz screen, will probably take a while to finish.
Just wondering why my RaspiBlitz doesn't have the default localhost associated. I just installed the V1.0 image, then updated to V1.1 and never touched the localhost configuration. Maybe these additional localhost parameters should be added by default to the script?

[openoms]

Ok, glad to hear that it is working now, but it will still take some time to index. If it freezes a restart should not be a problem, it can continue indexing the blocks where it left off.
I don`t know why the localhost parameters did not appear on your setup, those are the default for electrs so should not be needed to be added on. (see in nano $HOME/electrs/src/config.rs)

The script has the source $HOME/.cargo/env which runs: export PATH="$HOME/.cargo/bin:$PATH" and we are using the full path of for cargo run or running from the directory as you did, so updating the PATH is not necessary.

I just installed the V1.0 image, then updated to V1.1

There must be something unique about your setup.
Did you not burn the SDcard fresh to update to v1.1?

[Himbeergeld]

It has now finished indexing without any problem, thanks again! It's now at a point where it is frequently printing:

DEBUG - downloading new block headers (569769 already indexed) from ....
TRACE - downloaded 1 block headers
INFO - best=000.. height=569769 @ 2019-04-01T18:58:56Z (1 left to index)
DEBUG - applying 1 new headers from height 569769

So I guess it is just indexing each new block. So it means the script execution is basically finished. But what is the next step? Can I cancel this and reboot? Any recommendation?

Regarding my setup, yes, I burned a fresh SD card with v1.1 and I didn't modify anything in the configurations. I'm running RTL but I would assume this has no impact on localhost settings?!

[openoms]

Yes, the indexing is done and it is just keeping up to date with the blocks. It runs in the foreground this way, so if you close the terminal the service stops.
You can exit now and just run electrs again when you need it. The indexing should take only a couple of minutes even after being off for days.

Alternatively continue with this step for automatic start and SSL connection:
https://github.com/openoms/bitcoin-tutorials/blob/master/electrs/README.md#set-up-the-electrs-service-autostart-and-the-connection-over-ssl

[openoms]

I'm running RTL but I would assume this has no impact on localhost settings?!

No it should not. Did not see this problem before, will keep an eye. You have the extra options you need now, if it would reappear.

[Himbeergeld]

I probably underestimated the implications of the next steps. I don't have dynamic DNS and my RaspiBlitz is in the guest network of my router, where port-forwarding is not possible. Looks like I'm stuck at this point. Initially my plan was to run a Lightning wallet just locally, inside the guest network of my router. Do you think there is any way to achieve this without dynamic DNS or port forwarding? Happy for any advice.

[openoms]

So now you electrs is runni g and you should be able to access it within your local LAN with the Electrum wallet ,but not Eclair.
Eclair is a self sufficient Lightning node which uses Electrum servers to access the Bitcoin nodes. You can use Eclair, but not with your own Electrum server if you don't want to do port forwarding and dynDNS.

You can go around port forwarding using Tor.
Take a look at this guide: https://github.com/openoms/bitcoin-tutorials/blob/master/Zeus_to_RaspiBlitz_through_Tor.md
The Zeus wallet can be a remote control for your Lightning node on the RaspiBlitz.

It is planned to have this guides set up to be available from the menu as an automated option in a future release.
You can already activate Tor from the service menu.

[openoms]

I wonder how are things working out for you @Himbeergeld ?

[Himbeergeld]

Thanks for asking, still in progress, so many bumps in the way of the Electrum installation on my Debian 9 (e.g. haven't even installed sudo, etc. ). As I'm a Linux beginner each issue takes some time for me to research. I will let you know when successful.

[Himbeergeld]

Currently the script complains that "Electrum requires Python version >= 3.6.1" even though it shows python3-pip version 9.0.1-2. But don't worry, I'll research myself, it's the best way of learning. :-)

[openoms]

:) that is about your Python3 version, not PIP.
Update with: sudo apt install python3

To update every package:
sudo apt update
sudo apt upgrade

[Himbeergeld]

sudo apt install python3 responds
python3 is already the newest version (3.5.3-1).
sudo apt update and sudo apt upgrade also doesn't change the situation.
Also tried following the procedure on Electrum website but at the end with your script or Electrum website procedure I always get this:

Processing ./Electrum-3.3.4.tar.gz
Complete output from command python setup.py egg_info:
Error: Electrum requires Python version >= 3.6.1...
----------------------------------------
Command "python setup.py egg_info" failed with error code 1 in ...

I admit this issue is drifting away from the original question. Fine for me if you close it but if you have any idea, would be great.