5. Swarm for Node-Operators¶
This section is about how to run your swarm node, or deploy a seperate private or public swarm network.
5.1. Installation and Updates¶
Swarm runs on all major platforms (Linux, macOS, Windows, Raspberry Pi, Android, iOS).
Swarm was written in golang and requires the go-ethereum client geth to run.
The swarm package has not been extensively tested on platforms other than Linux and macOS.
5.2. Download pre-compiled Swarm binaries¶
Pre-compiled binaries for Linux, macOS and Windows are available to download via our official homepage.
5.3. Setting up Swarm in Docker¶
You can run it with optional arguments, e.g.:
$ docker run -it ethersphere/swarm --debug --verbosity 4
In order to up/download, you need to expose the HTTP api port (here: to localhost:8501) and set the HTTP address:
$ docker run -p 8501:8500/tcp -it ethersphere/swarm --httpaddr=0.0.0.0 --debug --verbosity 4
In this example, you can use
swarm --bzzapi http://localhost:8501 up testfile.md to upload
testfile.md to swarm using the Docker node, and you can get it back e.g. with
Note that if you want to use a pprof HTTP server, you need to expose the ports and set the address (with
$ docker run -p 8501:8500/tcp -v /tmp/hostdata:/data -it --name swarm1 ethersphere/swarm --httpaddr=0.0.0.0 --datadir /data --debug --verbosity 4
Then, you can attach the console with:
$ docker exec -it swarm1 geth attach /data/bzzd.ipc
You can also open a terminal session inside the container:
$ docker exec -it swarm1 /bin/sh
5.4. Installing Swarm from source¶
The Swarm source code for can be found on https://github.com/ethersphere/swarm
5.4.1. Prerequisites: Go and Git¶
Building the Swarm binary requires the following packages:
Grab the relevant prerequisites and build from source.
5.4.2. Configuring the Go environment¶
You should then prepare your Go environment.
5.4.3. Download and install Geth¶
Once all prerequisites are met, download and install Geth from https://github.com/ethereum/go-ethereum
5.4.4. Compiling and installing Swarm¶
Once all prerequisites are met, and you have
geth on your system, clone the Swarm git repo and build from source:
$ git clone https://github.com/ethersphere/swarm $ cd swarm $ make swarm
Alternatively you could also use the Go tooling and download and compile Swarm from master via:
$ go get -d github.com/ethersphere/swarm $ go install github.com/ethersphere/swarm/cmd/swarm
You can now run
swarm to start your Swarm node.
Let’s check if the installation of
swarm was successful:
PATH is not set and the
swarm command cannot be found, try:
$ $GOPATH/bin/swarm version
This should return some relevant information. For example:
Swarm Version: 0.3 Network Id: 0 Go Version: go1.10.1 OS: linux GOPATH=/home/user/go GOROOT=/usr/local/go
5.4.5. Updating your client¶
To update your client simply download the newest source code and recompile.
5.5. Running the Swarm Go-Client¶
To start a basic Swarm node you must have both
swarm installed on your machine. You can find the relevant instructions in the Installation and Updates section.
geth is the go-ethereum client, you can read up on it in the Ethereum Homestead documentation.
To start Swarm you need an Ethereum account. You can create a new account in
geth by running the following command:
$ geth account new
You will be prompted for a password:
Your new account is locked with a password. Please give a password. Do not forget this password. Passphrase: Repeat passphrase:
Once you have specified the password, the output will be the Ethereum address representing that account. For example:
Using this account, connect to Swarm with
$ swarm --bzzaccount <your-account-here> # in our example $ swarm --bzzaccount 2f1cd699b0bf461dcfbf0098ad8f5587b038f0f1
(You should replace
2f1cd699b0bf461dcfbf0098ad8f5587b038f0f1 with your account address key).
Remember your password. There is no forgot my password option for
5.6. Interacting with Swarm¶
$BZZKEY$ refers to your account address key.
Swarm is fully compatible with Geth Console commands. For example, you can list your peers using
admin.peers, add a peer using
admin.addPeer, and so on.
You can use Swarm with CLI flags and environment variables. See a full list in the Configuration .
5.7. How do I enable ENS name resolution?¶
The Ethereum Name Service (ENS) is the Ethereum equivalent of DNS in the classic web. It is based on a suite of smart contracts running on the Ethereum mainnet.
In order to use ENS to resolve names to swarm content hashes,
swarm has to connect to a
geth instance that is connected to the Ethereum mainnet. This is done using the
First you must start your geth node and establish connection with Ethereum main network with the following command:
for a full geth node, or
$ geth --syncmode=light
for light client mode.
Syncing might take a while. When you use the light mode, you don’t have to sync the node before it can be used to answer ENS queries. However, please note that light mode is still an experimental feature.
After the connection is established, open another terminal window and connect to Swarm:
Verify that this was successful by pointing your browser to http://localhost:8500/bzz:/theswarm.eth/
5.7.1. Using Swarm together with the testnet ENS¶
It is also possible to use the Ropsten ENS test registrar for name resolution instead of the Ethereum main .eth ENS on mainnet.
Run a geth node connected to the Ropsten testnet
$ geth --testnet
Then launch the
swarm; connecting it to the geth node (
Swarm will automatically use the ENS deployed on Ropsten.
For other ethereum blockchains and other deployments of the ENS contracts, you can specify the contract addresses manually. For example the following command:
$ swarm --ens-api eth:<contract 1>@/home/user/.ethereum/geth.ipc \ --ens-api test:<contract 2>@ws:<address 1> \ --ens-api <contract 3>@ws:<address 2>
Will use the
geth.ipc to resolve
.eth names using the contract at address
<contract 1> and it will use
ws:<address 1> to resolve
.test names using the contract at address
<contract 2>. For all other names it will use the ENS contract at address
<contract 3> on
5.7.2. Using an external ENS source¶
Take care when using external sources of information. By doing so you are trusting someone else to be truthful. Using an external ENS source may make you vulnerable to man-in-the-middle attacks. It is only recommended for test and development environments.
Maintaining a fully synced Ethereum node comes with certain hardware and bandwidth constraints, and can be tricky to achieve. Also, light client mode, where syncing is not necessary, is still experimental.
An alternative solution for development purposes is to connect to an external node that you trust, and that offers the necessary functionality through HTTP.
If the external node is running on IP 220.127.116.11 port 8545, the command would be:
$ swarm --ens-api http://18.104.22.168:8545
You can also use
https. But keep in mind that Swarm does not validate the certificate.
5.8. Alternative modes¶
Below are examples on ways to run
5.8.1. Swarm in singleton mode (no peers)¶
If you don’t want your swarm node to connect to any existing networks, you can provide it with a custom network identifier using
--bzznetworkid with a random large number.
5.8.2. Adding enodes manually¶
By default, Swarm will automatically seek out peers in the network.
Additionally you can manually start off the connection process by adding one or more peers using the
admin.addPeer console command.
(You can also do this in the Geth Console, as seen in Section 3.2.)
When you stop a node, all peer connections will be saved. When you start again, the node will try to reconnect to those peers automatically.
Where ENODE is the enode record of a swarm node. Such a record looks like the following:
The enode of your swarm node can be accessed using
geth connected to
geth is used for two different purposes here: You use it to run an Ethereum Mainnet node for ENS lookups. But you also use it to “attach” to the Swarm node to send commands to it.
5.8.3. Connecting to the public Swarm cluster¶
By default Swarm connects to the public Swarm testnet operated by the Ethereum Foundation and other contributors.
The nodes the team maintains function as a free-to-use public access gateway to Swarm, so that users can experiment with Swarm without the need to run a local node. To download data through the gateway use the
5.8.4. Metrics reporting¶
Swarm uses the go-metrics library for metrics collection. You can set your node to collect metrics and push them to an influxdb database (called metrics by default) with the default settings. Tracing is also supported. An example of a default configuration is given below:
$ swarm --bzzaccount <bzzkey> \ --debug \ --metrics \ --metrics.influxdb.export \ --metrics.influxdb.endpoint "http://localhost:8086" \ --metrics.influxdb.username "user" \ --metrics.influxdb.password "pass" \ --metrics.influxdb.database "metrics" \ --metrics.influxdb.host.tag "localhost" \ --verbosity 4 \ --tracing \ --tracing.endpoint=jaeger:6831 \ --tracing.svc myswarm
5.9. Go-Client Command line options Configuration¶
swarm executable supports the following configuration options:
- Configuration file
- Environment variables
- Command line
Options provided via command line override options from the environment variables, which will override options in the config file. If an option is not explicitly provided, a default will be chosen.
In order to keep the set of flags and variables manageable, only a subset of all available configuration options are available via command line and environment variables. Some are only available through a TOML configuration file.
Swarm reuses code from ethereum, specifically some p2p networking protocol and other common parts. To this end, it accepts a number of environment variables which are actually from the
geth environment. Refer to the geth documentation for reference on these flags.
This is the list of flags inherited from
--identity --bootnodes --datadir --keystore --port --nodiscover --v5disc --netrestrict --nodekey --nodekeyhex --maxpeers --nat --ipcdisable --ipcpath --password
5.10. Config File¶
swarm can be executed with the
dumpconfig command, which prints a default configuration to STDOUT, and thus can be redirected to a file as a template for the config file.
A TOML configuration file is organized in sections. The below list of available configuration options is organized according to these sections. The sections correspond to Go modules, so need to be respected in order for file configuration to work properly. See https://github.com/naoina/toml for the TOML parser and encoder library for Golang, and https://github.com/toml-lang/toml for further information on TOML.
To run Swarm with a config file, use:
$ swarm --config /path/to/config/file.toml
5.11. General configuration parameters¶
|Config file||Command-line flag||Environment variable||Default value||Description|
|n/a||–config||n/a||n/a||Path to config file in TOML format|
|n/a||–bzzapi||n/a||http://127.0.0.1:8500||Swarm HTTP endpoint|
|BzzAccount||–bzzaccount||SWARM_ACCOUNT||Swarm account key|
|BzzKey||n/a||n/a||n/a||Swarm node base address (\(hash(PublicKey)hash(PublicKey))\). This is used to decide storage based on radius and routing by kademlia.|
|Cors||–corsdomain||SWARM_CORS||Domain on which to send Access-Control-Allow-Origin header (multiple domains can be supplied separated by a ‘,’)|
|n/a||–debug||n/a||n/a||Prepends log messages with call-site location (file and line number)|
|n/a||–defaultpath||n/a||n/a||path to file served for empty url path (none)|
|n/a||–delivery-skip-check||SWARM_DELIVERY_SKIP_CHECK||false||Skip chunk delivery check (default false)|
|EnsApi||–ens-api||SWARM_ENS_API||<$GETH_DATADIR>/geth.ipc||Ethereum Name Service API address|
|EnsRoot||–ens-addr||SWARM_ENS_ADDR||ens.TestNetAddress||Ethereum Name Service contract address|
|ListenAddr||–httpaddr||SWARM_LISTEN_ADDR||127.0.0.1||Swarm listen address|
|n/a||–manifest value||n/a||true||Automatic manifest upload (default true)|
|n/a||–mime value||n/a||n/a||Force mime type on upload|
|Path||–datadir||GETH_DATADIR||<$GETH_DATADIR>/swarm||Path to the geth configuration directory|
|Port||–bzzport||SWARM_PORT||8500||Port to run the http proxy server|
|PublicKey||n/a||n/a||n/a||Public key of swarm base account|
|n/a||–recursive||n/a||false||Upload directories recursively (default false)|
|n/a||–stdin||n/a||Reads data to be uploaded from stdin|
|n/a||–store.path value||SWARM_STORE_PATH||<$GETH_ENV_DIR>/swarm/bzz-<$BZZ_KEY>/chunks||Path to leveldb chunk DB|
|n/a||–store.size value||SWARM_STORE_CAPACITY||5000000||Number of chunks (5M is roughly 20-25GB) (default 5000000)]|
|n/a||–store.cache.size value||SWARM_STORE_CACHE_CAPACITY||5000||Number of recent chunks cached in memory (default 5000)|
|n/a||–sync-update-delay value||SWARM_ENV_SYNC_UPDATE_DELAY||Duration for sync subscriptions update after no new peers are added (default 15s)|
|SyncDisabled||–nosync||SWARM_ENV_SYNC_DISABLE||false||Disable Swarm node synchronization|
|SwapBackendURL||–swap-backend-url||SWARM_SWAP_BACKEND_URL||URL of the Ethereum API provider to use to settle SWAP payments|
|SwapPaymentThreshold||–swap-payment-threshold||SWARM_SWAP_PAYMENT_THRESHOLD||1000000||honey amount at which payment is triggered|
|SwapDisconnectThreshold||–swap-disconnect-threshold||SWARM_SWAP_DISCONNECT_THRESHOLD||1500000||honey amount at which a peer disconnects|
|SwapInitialDeposit||–swap-initial-deposit||SWARM_INITIAL_DEPOSIT||0||Initial deposit amount for swap chequebook|
|SwapLogPath||–swap-audit-logpath||SWARM_SWAP_LOG_PATH||n/a||Write execution logs of swap audit to the given directory|
|Contract||–swap-chequebook||SWARM_CHEQUEBOOK_ADDR||0x0||Swap chequebook contract address|
|n/a||–verbosity value||n/a||3||Logging verbosity: 0=silent, 1=error, 2=warn, 3=info, 4=debug, 5=detail|
|n/a||–ws||n/a||false||Enable the WS-RPC server|
|n/a||–wsaddr value||n/a||localhost||WS-RPC server listening interface|
|n/a||–wsport value||n/a||8546||WS-RPC server listening port|
|n/a||–wsapi value||n/a||n/a||API’s offered over the WS-RPC interface|
|n/a||–wsorigins value||n/a||n/a||Origins from which to accept websockets requests|
|n/a||n/a||SWARM_AUTO_DEFAULTPATH||false||Toggle automatic manifest default path on recursive uploads (looks for index.html)|