Files
xrpl-dev-portal/content/tutorials/manage-the-rippled-server/install-rippled.md
Jennifer Hasegawa 9d0510a212 feedback
2018-11-01 17:31:04 -07:00

20 KiB

Install rippled

For production, you can install a rippled instance using Ripple's binary executable for Ubuntu or CentOS/Red Hat.

For development, you can install a rippled instance using Ripple's binary executable for macOS, or you can build and run rippled from source for Ubuntu.

For installation information for other platforms, see the rippled repository.

Minimum System Requirements

A rippled server should run comfortably on commodity hardware, to make it inexpensive to participate in the network. At present, we recommend the following minimum requirements:

  • Operating System:
    • Production: CentOS or RedHat Enterprise Linux (latest release) or Ubuntu (16.04+) supported
    • Development: Mac OS X, Windows (64-bit), or most Linux distributions
  • CPU: 64-bit x86_64, 2+ cores
  • Disk: Minimum 50GB SSD recommended (1000 IOPS, more is better) for the database partition
  • RAM:
    • Testing: 8GB+
    • Production: 32 GB

Amazon EC2's m3.large VM size may be appropriate depending on your workload. A fast network connection is preferable. Any increase in a server's client-handling load increases resources needs.

Tip: For recommendations beyond the minimum requirements, see Capacity Planning.

Installation on Ubuntu with alien

This section assumes that you are using Ubuntu 15.04 or later.

  1. Install yum-utils and alien:

     $ sudo apt-get update
     $ sudo apt-get install yum-utils alien
    
  2. Install the Ripple RPM repository:

     $ sudo rpm -Uvh https://mirrors.ripple.com/ripple-repo-el7.rpm
    
  3. Download the rippled software package:

     $ yumdownloader --enablerepo=ripple-stable --releasever=el7 rippled
    
  4. Verify the signature on the rippled software package:

     $ sudo rpm --import https://mirrors.ripple.com/rpm/RPM-GPG-KEY-ripple-release && rpm -K rippled*.rpm
    
  5. Install the rippled software package:

     $ sudo alien -i --scripts rippled*.rpm && rm rippled*.rpm
    
  6. Configure the rippled service to start on system boot:

     $ sudo systemctl enable rippled.service
    
  7. Start the rippled service

     $ sudo systemctl start rippled.service
    

For next steps, see Postinstall and Additional Configuration.

Installation on CentOS/Red Hat with yum

This section assumes that you are using CentOS 7 or Red Hat Enterprise Linux 7.

  1. Install the Ripple RPM repository:

     $ sudo rpm -Uvh https://mirrors.ripple.com/ripple-repo-el7.rpm
    
  2. Install the rippled software package:

     $ sudo yum install --enablerepo=ripple-stable rippled
    
  3. Configure the rippled service to start on system boot:

     $ sudo systemctl enable rippled.service
    
  4. Start the rippled service

     $ sudo systemctl start rippled.service
    

For next steps, see Postinstall and Additional Configuration.

Installation on macOS

At this time, we don't recommend using the macOS platform for rippled production use. For production, consider using the Ubuntu platform, which has received the highest level of quality assurance and testing.

That said, macOS is suitable for many development and testing tasks.

This section assumes that you are using macOS X 10.0 or higher. TODO: okay?

{% set n = cycler(* range(1,99)) %}

{{n.next()}}. Install Xcode.

{{n.next()}}. Install Xcode command line tools.

    xcode-select --install

{{n.next()}}. Install Homebrew.

    ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

{{n.next()}}. Update Homebrew.

    brew update

{{n.next()}}. Use Homebrew to install dependencies.

    brew install git cmake pkg-config protobuf openssl ninja

{{n.next()}}. Install a supported version of Boost. Supported versions include Boost 1.67.x - 1.68.x. TODO: correctly stated supported versions? In the rippled repo, we say "Boost 1.67 or later is required." <-- Based on our discussion in this doc, I don't think this open-ended "later" part is true, is it?

Pick one of the following Boost installation methods: ***TODO: I provided the two options below because these are the most straightforward ways for a user to install one of the two supported versions of Boost. In the future, if there are more supported versions or more brew boost formalae, there may be more options to document.***

* To install Boost 1.67.0, use `brew install boost@1.67`. ***TODO: When Boost 1.67.0 is no longer supported, we'll need to change this formula. Until then, the user will be able to use this command to get a supported version of Boost. `brew install boost` also installs 1.67.0, but as discussed, we don't want to provide the command because it won't always install 1.67.0 and may cause the user to install an unsupported version. There is no brew command to install 1.68 right now.***

* To install Boost 1.68.0, compile Boost yourself. ***TODO: When the supported versions change above, we can also change these instructions, if needed. It would be great to genericize - but the way that the Boost brew and compiles are working right now - being version-specific seems necessary to keep folks from inadvertently installing an unsupported version.***

    1. Download [Boost 1.68.0](https://dl.bintray.com/boostorg/release/1.68.0/source/boost_1_68_0.tar.bz2) to an appropriate directory. ***TODO: okay to do it outside of CLI? If we want to give them CLI instructions, need to have them run `brew install wget` and then `wget https://dl.bintray.com/boostorg/release/1.68.0/source/boost_1_68_0.tar.bz2`.***

    2. Extract `boost_1_68_0.tar.bz2`.

          `tar xvzf boost_1_68_0.tar.bz2`

    3. Change to the newly created `boost_1_68_0` directory.

          `cd boost_1_68_0`

    4. Prepare the Boost.Build system for use.

          `./bootstrap.sh`

    5. Build the separately-compiled Boost libraries. This may take about 10 minutes, depending on your hardware specs. 11:14

          `./b2 cxxflags="-std=c++14" -j 4`

        **Tip:** This example uses 4 processes to build in parallel. The best number of processes to use depends on how many CPU cores your hardware has available. Use `sysctl -n hw.ncpu` to get your CPU core count.

{{n.next()}}. Ensure that the BOOST_ROOT environment variable points to the directory created by your Boost installation. Put this environment variable in your .bash_profile file so it's automatically set when you log in.

* If you used Homebrew to install Boost, ensure that `BOOST_ROOT` is set to the location where Homebrew installed Boost. For example:

    `export BOOST_ROOT=/usr/local/Cellar/boost/1.67.0_1`

    To find your Boost install directory, use `brew info boost`.

* If you manually compiled and installed Boost, ensure that `BOOST_ROOT` is set to the location where you installed Boost. For example:

    `export BOOST_ROOT=/Users/tblee/boost/boost_1_68_0`

{{n.next()}}. If you updated your .bash_profile file in the previous step, be sure to source it. For example:

    source .bash_profile

{{n.next()}}. Clone the rippled source code into your desired location and access the rippled directory. To do this, you'll need to set up Git (installed earlier using Homebrew) and GitHub. For example, you'll need to create a GitHub account and set up your SSH key. For more information, see Set up git. TODO: does this sound okay for introducing what a user may have to do to set up git and GitHub?

    git clone git@github.com:ripple/rippled.git
    cd rippled

{{n.next()}}. By default, cloning puts you on the develop branch. Use this branch if you are doing development work and want the latest set of untested features.

  If you want the latest stable release, checkout the `master` branch.

    git checkout master

  Or, you can checkout one of the tagged releases listed on [GitHub](https://github.com/ripple/rippled/releases).

  If you want to test out the latest release candidate, checkout the `release` branch:

    git checkout release

{{n.next()}}. In the rippled directory you just cloned, create your build directory and access it. TODO: Does this "In the rippled directory you just cloned" language make sense? I just want to tell the user where to create the build directory. For example:

    mkdir my_build
    cd my_build

{{n.next()}}. Build rippled. This could take about 5 minutes, depending on your hardware specs.

    cmake -G "Unix Makefiles" -DCMAKE_BUILD_TYPE=Debug ..

  You can set `CMAKE_BUILD_TYPE` to the `Debug` or `Release` build type. All four standard [`CMAKE_BUILD_TYPE`](https://cmake.org/cmake/help/v3.0/variable/CMAKE_BUILD_TYPE.html) values are supported.

{{n.next()}}. Run the build using CMake. This could take about 10 minutes, depending on your hardware specs.

    cmake --build . -- -j 4

  This example uses a `-j` parameter set to `4`, which uses four processes to build in parallel. The best number of processes to use depends on how many CPU cores your hardware has available. Use `sysctl -n hw.ncpu` to get your CPU core count.

{{n.next()}}. Run unit tests built into the server executable. This could take about 5 minutes, depending on your hardware specs. (optional, but recommended)

    ./rippled --unittest

{{n.next()}}. rippled requires the rippled.cfg configuration file to run. You can find an example config file, rippled-example.cfg in rippled/cfg. Make a copy and save it as rippled.cfg in a location that enables you to run rippled as a non-root user. Access the rippled directory and run:

    mkdir -p /etc/opt/ripple ***TODO: changed this from ~/.config/ripple to match guidance coming in conf-file-location.md PR -- okay?***
    cp cfg/rippled-example.cfg /etc/opt/ripple/rippled.cfg

{{n.next()}}. Edit rippled.cfg to set necessary file paths. The user you plan to run rippled as must have write permissions to all of the paths you specify here. TODO: I think we need to add this step to the centos and ubuntu tasks, correct? For centos and ubuntu, the config files are automatically placed in the correct directories, but the following paths still need to be manually updated to avoid permission issues, correct?

  * Set the `[node_db]` path to the location where you want to store the ledger database.

  * Set the `[database_path]` to the location where you want to store other database data. (This includes an SQLite database with configuration data, and is typically one level above the `[node_db]` path field.)

  * Set the `[debug_logfile]` to a path where `rippled` can write logging information.

  These are the only configurations required for `rippled` to start up successfully. All other configuration is optional and can be tweaked after you have a working server. For more information, see [Additional Configurations](#additional-configuration).

{{n.next()}}. rippled requires the validators.txt file to run. You can find an example validators file, validators-example.txt, in rippled/cfg/. Make a copy and save it as validators.txt in the same folder as your rippled.cfg file. Access the rippled directory and run:

    cp cfg/validators-example.txt /etc/opt/ripple ***TODO: changed this from ~/.config/ripple to match guidance coming in conf-file-location.md PR -- okay?***

  **Warning:** Ripple has designed a decentralization plan with maximum safety in mind. During the transition, you should not modify the `validators.txt` file except as recommended by Ripple. Even minor modifications to your validator settings could cause your server to diverge from the rest of the network and report out of date, incomplete, or inaccurate data. Acting on such data can cause you to lose money.

{{n.next()}}. Access your build directory, my_build for example, and start the rippled service.

    sudo ./rippled

For next steps, see Postinstall and Additional Configuration.

Postinstall

It can take several minutes for rippled to sync with the rest of the network, during which time it outputs warnings about missing ledgers.

Here's an excerpt of what you can expect to see in your terminal: TODO: you'll see this in the CLI for the macOS install because we tell them to start the service using sudo ./rippled, but not for centos and ubuntu. For centos and ubuntu, we tell the user to start the service using sudo systemctl start rippled.service, which starts the service, but doesn't display a log of the service starting up. At least for first-time start up - would it make sense to give the user this start up command (sudo ./rippled) instead so that they can see it running? Perhaps in Postinstall we can give them the sudo systemctl start rippled.service command for when they want to start the service, but don't need to see the running log?

    2018-Oct-26 18:21:39.593738 JobQueue:NFO Auto-tuning to 6 validation/transaction/proposal threads.
    2018-Oct-26 18:21:39.599634 Amendments:DBG Amendment 4C97EBA926031A7CF7D7B36FDE3ED66DDA5421192D63DE53FFB46E43B9DC8373 is supported.
    2018-Oct-26 18:21:39.599874 Amendments:DBG Amendment 6781F8368C4771B83E8B821D88F580202BCB4228075297B19E4FDC5233F1EFDC is supported.
    2018-Oct-26 18:21:39.599965 Amendments:DBG Amendment 42426C4D4F1009EE67080A9B7965B44656D7714D104A72F9B4369F97ABF044EE is supported.
    2018-Oct-26 18:21:39.600024 Amendments:DBG Amendment 08DE7D96082187F6E6578530258C77FAABABE4C20474BDB82F04B021F1A68647 is supported.
    ...
    2018-Oct-26 18:21:39.603201 OrderBookDB:DBG Advancing from 0 to 3
    2018-Oct-26 18:21:39.603291 OrderBookDB:DBG OrderBookDB::update>
    2018-Oct-26 18:21:39.603480 OrderBookDB:DBG OrderBookDB::update< 0 books found
    2018-Oct-26 18:21:39.649617 ValidatorList:DBG Loading configured trusted validator list publisher keys
    2018-Oct-26 18:21:39.649709 ValidatorList:DBG Loaded 0 keys
    2018-Oct-26 18:21:39.649798 ValidatorList:DBG Loading configured validator keys
    2018-Oct-26 18:21:39.650213 ValidatorList:DBG Loaded 5 entries
    2018-Oct-26 18:21:39.650266 ValidatorSite:DBG Loading configured validator list sites
    2018-Oct-26 18:21:39.650319 ValidatorSite:DBG Loaded 0 sites
    2018-Oct-26 18:21:39.650829 NodeObject:DBG NodeStore.main target size set to 131072
    2018-Oct-26 18:21:39.650876 NodeObject:DBG NodeStore.main target age set to 120000000000
    2018-Oct-26 18:21:39.650931 TaggedCache:DBG LedgerCache target size set to 256
    2018-Oct-26 18:21:39.650981 TaggedCache:DBG LedgerCache target age set to 180000000000
    2018-Oct-26 18:21:39.654252 TaggedCache:DBG TreeNodeCache target size set to 512000
    2018-Oct-26 18:21:39.654336 TaggedCache:DBG TreeNodeCache target age set to 90000000000
    2018-Oct-26 18:21:39.674131 NetworkOPs:NFO Consensus time for #3 with LCL AF8D8984A226AE7099D8A9749B09CE1D84360D5AF9FB86CE2F37500FE1009F9D
    2018-Oct-26 18:21:39.674271 ValidatorList:DBG 5  of 5 listed validators eligible for inclusion in the trusted set
    2018-Oct-26 18:21:39.674334 ValidatorList:DBG Using quorum of 4 for new set of 5 trusted validators (5 added, 0 removed)
    2018-Oct-26 18:21:39.674400 LedgerConsensus:NFO Entering consensus process, watching, synced=no
    2018-Oct-26 18:21:39.674475 LedgerConsensus:NFO Consensus mode change before=observing, after=observing
    2018-Oct-26 18:21:39.674539 NetworkOPs:DBG Initiating consensus engine
    2018-Oct-26 18:21:39.751225 Server:NFO Opened 'port_rpc_admin_local' (ip=127.0.0.1:5005, admin IPs:127.0.0.1, http)
    2018-Oct-26 18:21:39.751515 Server:NFO Opened 'port_peer' (ip=0.0.0.0:51235, peer)
    2018-Oct-26 18:21:39.751689 Server:NFO Opened 'port_ws_admin_local' (ip=127.0.0.1:6006, admin IPs:127.0.0.1, ws)
    2018-Oct-26 18:21:39.751915 Application:FTL Startup RPC:
    {
    	"command" : "log_level",
    	"severity" : "warning"
    }


    2018-Oct-26 18:21:39.752079 Application:FTL Result: {}


    2018-Oct-26 18:22:33.013409 NetworkOPs:WRN We are not running on the consensus ledger
    2018-Oct-26 18:22:33.013875 LedgerConsensus:WRN Need consensus ledger 81804C95ADE119CC874572BAF24DB0C0D240AC58168597951B0CB64C4DA2C628
    2018-Oct-26 18:22:33.883648 LedgerConsensus:WRN View of consensus changed during open status=open,  mode=wrongLedger
    2018-Oct-26 18:22:33.883815 LedgerConsensus:WRN 81804C95ADE119CC874572BAF24DB0C0D240AC58168597951B0CB64C4DA2C628 to 9250C6C8326A48C339E6F99167F4E6BFD0DB00C35518027724D7B376340D21A1
    2018-Oct-26 18:22:33.884068 LedgerConsensus:WRN {"accepted":true,"account_hash":"BBA0E7273005D42E5548DD6456E5AD1F7C89B6EDCB01881E1EECD393E8545947","close_flags":0,"close_time":593893350,"close_time_human":"2018-Oct-26 18:22:30.000000","close_time_resolution":30,"closed":true,"hash":"9250C6C8326A48C339E6F99167F4E6BFD0DB00C35518027724D7B376340D21A1","ledger_hash":"9250C6C8326A48C339E6F99167F4E6BFD0DB00C35518027724D7B376340D21A1","ledger_index":"3","parent_close_time":593893290,"parent_hash":"AF8D8984A226AE7099D8A9749B09CE1D84360D5AF9FB86CE2F37500FE1009F9D","seqNum":"3","totalCoins":"100000000000000000","total_coins":"100000000000000000","transaction_hash":"0000000000000000000000000000000000000000000000000000000000000000"}
    2018-Oct-26 18:23:03.034119 InboundLedger:WRN Want: D901E53926E68EFDA33172DDAC74E8C767D280B68EE68E3010AB0E3179D07B1C
    2018-Oct-26 18:23:03.034334 InboundLedger:WRN Want: 1C01EE79083DE5CE76F3634519D6364C589C4D48631CB9CD10FC2408F87684E2
    2018-Oct-26 18:23:03.034560 InboundLedger:WRN Want: 8CFE3912001BDC5B2C4B2691F3C7811B9F3F193E835D293459D80FBF1C4E684E
    2018-Oct-26 18:23:03.034750 InboundLedger:WRN Want: 8DFAD21AD3090DE5D6F7592B3821C3DA77A72287705B4CF98DC0F84D5DD2BDF8

Once your rippled has synchronized with the rest of the network, you have a fully functional stock rippled server that you can use for local signing and API access to the XRP Ledger. Use rippled server states to tell whether your rippled server has synchronized with the network.

For information about communicating with your rippled server using the rippled API, see the rippled API reference.

Once you have your stock rippled server running, you may want to consider running it as a validating server. For information about validating servers and why you might want to run one, see Run rippled as a Validator.

Additional Configuration

rippled should connect to the XRP Ledger with the default configurations. However, you can change your settings by editing the rippled.cfg file. For recommendations about configuration settings, see Capacity Planning.

Changes to the [node_db], [debug_logfile], or [database_path] sections may require you to give the rippled user and group ownership to your new configured path: TODO: I added [node_db] to this list - okay? When I try the following command, chown -R rippled:rippled /var/lib/rippled/db/rocksdb, I get chown: rippled: illegal group name. Is rippled:rippled just an example below -- or should there be a rippled group that is created by the build/install process?

$ chown -R rippled:rippled <configured path>

Restart rippled for any configuration changes to take effect: TODO: when I try this on macOS, I get sudo: service: command not found. I don't think the service command exists for macOS. For macOS, can the user just stop (ctrl+c? Not sure if that is the proper way to stop the service) and start rippled (sudo ./rippled) instead? For centos and ubuntu, we use sudo systemctl start rippled.service to start the service - below should we use sudo systemctl restart rippled.service - just to be consistent?

$ sudo service rippled restart