Skip to content

Latest commit

 

History

History
579 lines (529 loc) · 20 KB

MAC_M1.md

File metadata and controls

579 lines (529 loc) · 20 KB

Mac M1 and M2 (Apple Silicon) Installation

<< Back

How to verify that you have an Apple Silicon processor:

  1. Click the Apple icon in the top-left of the screen
  2. Click 'About This Mac'
  3. Look for the 'Chip' section at the top of the list
    1. Apple Silicon processors will begin with 'Apple'; as of 7/2023, these a variant of 'Apple M1' or 'Apple M2'

Frequently Asked Question:

  1. Why is this so complicated?

Apple Silicon processors use a different architecture (arm64/aarch64) than Intel processors (x86_64). Oracle, which is used for the VACOLS database, does not have binaries to run any of their database tools natively on arm64 for MacOS. Additionally, the Ruby gems therubyracer and jshint require the library v8@3.15, which can also only be compiled and installed on x86_64 processors. To work around this we use Rosetta to emulate x86_64 processors in the terminal, installing most of the Caseflow dependencies via the x86_64 version of Homebrew. It is important that while setting up your environment, you ensure you are in the correct terminal type and in the correct directory so that you do not install or compile a dependency with the wrong architecture.

  1. I am running into errors! Where can I go for help?

See the Installation Workarounds section for common or previously relevant workarounds that may help. Additionally, join the #bid_appeals_mac_support channel in Slack (or ask your scrum master to add you). You can search that channel to see if your issue has been previously discussed or post what step you are having a problem on and what you've done so far.

Ensure command line tools are installed via Self Service Portal prior to starting

Follow these instructions as closely as possible. If a folder is specified for running terminal commands, ensure you are in that directory prior to running the command(s). If you can't complete a step, ask for help in the #bid_appeals_mac_support channel of the Benefits Integrated Delivery (BID) Slack workspace.

Clone these Repos

  1. Create a ~/dev/appeals/ directory

  2. Clone the following repo using git clone into this directory

  3. Optional for setting up a machine, though you may work with these in the future

  4. If you cannot clone the above, you might need to do this setup

Homebrew Installation

  1. Install homebrew from self-service portal

Docker Installation

Note: We do not use Docker Desktop due to licensing. We recommend using Colima to run the docker service.

  1. Open terminal and run:
    1. brew install docker docker-compose colima
    2. mkdir -p ~/.docker/cli-plugins
    3. ln -sfn /opt/homebrew/opt/docker-compose/bin/docker-compose ~/.docker/cli-plugins/docker-compose

Install UTM and VACOLS VM

  1. Download UTM from this link
  2. Right click UTM.app and select “install with privilege management” then open the UTM.app
  3. Download the Vacols VM from this link
  4. After the file downloads, right click on it in “Finder” and select “Show Package Contents” and delete the view.plist file if it exists
  5. Right click on the application and select “Open With > UTM.app (default)”
  6. Select the “Play” button when it pops up in UTM
  7. Leave this running in the background. If you close the window, you can open it back up by repeating steps 5-7

Chromedriver, PDFtk Server, and wkhtmltox Installation

  1. Open terminal and run
    • brew install --cask chromedriver
  2. Once it successfully installs, run command
    • chromedriver –version
  3. There will be a pop up. Before clicking “OK,” navigate to System Settings > Privacy & Security
  4. Scroll down to the security section and it will say “chromedriver was blocked from use because it is not from an identified developer”
  5. Select “Allow Anyway”
  6. Select “Yes” from pop up
  7. Reopen terminal and once again run chromedriver –version
  8. Select “Open”
  9. Download and install from this link. When you receive a security warning, follow steps 3-6 for PDFtk server
  10. Download this file and move through the prompts. When you receive a security warning, follow steps 3-6 for wkhtmltox

Oracle “instantclient” Files

  1. Download these DMG files
  2. After downloading, double click on one of the folders and follow the instructions in INSTALL_IC_README.txt to copy the libraries

Postgres Download

  1. Download and install from this link

Configure x86_64 Homebrew

Run the below commands from your home directory

  1. In a terminal, create a homebrew directory under your home directory
    • mkdir homebrew
  2. In a terminal, run
    • curl -L https://github.com/Homebrew/brew/tarball/master | tar xz --strip 1 -C homebrew
  3. If you get a chdir error, run
    • mkdir homebrew && curl -L https://github.com/Homebrew/brew/tarball/master | tar xz --strip 1 -C homebrew

Rosetta

  1. Open standard terminal and run:
    • softwareupdate -–install-rosetta –-agree-to-license
  2. Once Rosetta is installed, find the default terminal in “Finder” > Applications
  3. Right click and select “Get Info”
  4. Select “Open using Rosetta”
    • Note: you can copy the standard terminal executable to your desktop and enable Rosetta on that, so that you don’t need to disable rosetta on the default terminal once Caseflow setup is complete

Modify your .zshrc File

  1. Run command open ~/.zshrc
  2. Add the following lines, if any of these are already set make sure to comment previous settings:
export PATH=~/homebrew/bin:${PATH}
eval "$(~/homebrew/bin/rbenv init -)"
eval "$(~/homebrew/bin/nodenv init -)"
eval "$(~/homebrew/bin/pyenv init --path)"

# Add Postgres environment variables for CaseFlow
export POSTGRES_HOST=localhost
export POSTGRES_USER=postgres
export POSTGRES_PASSWORD=postgres
export NLS_LANG=AMERICAN_AMERICA.UTF8
export FREEDESKTOP_MIME_TYPES_PATH=~/homebrew/share/mime/packages/freedesktop.org.xml export OCI_DIR=~/Downloads/instantclient_19_8
  1. Save file
  2. Go to your active terminal and enter source ~/.zshrc or create a new terminal Note: until rbenv, nodenv, and pyenv are installed, the eval commands will display a 'command not found' error when launching a terminal

Run dev setup scripts in Caseflow repo

VERY IMPORTANT NOTE: The below commands must be run in a Rosetta terminal until you reach the 'Running Caseflow' section

Script 1

  1. Open a new Rosetta terminal and ensure you are in the directory you cloned the Caseflow repo into (~/dev/appeals/caseflow) and run commands:
    1. git checkout grant/setup-m1
    2. ./scripts/dev_env_setup_step1.sh
    • If this fails, double check your .zshrc file to ensure your PATH has only the x86_64 brew

Note: If you run into errors installing any versions of openssl, see the "Installation Workarounds" section at the bottom of this document

Script 2

  1. In the Rosetta terminal, install pyenv and the required python2 version:
    1. brew install pyenv
    2. pyenv rehash
    3. pyenv install 2.7.18
    4. In the caseflow directory, run pyenv local 2.7.18 to set the version
  2. In the Rosetta terminal navigate to caseflow folder:
    1. run rbenv install $(cat .ruby-version)
    2. run rbenv rehash
    3. run gem install bundler -v $(grep -A 1 "BUNDLED WITH" Gemfile.lock | tail -n 1)
    4. run gem install pg:1.1.4 -- --with-pg-config=/Applications/Postgres.app/Contents/Versions/latest/bin/pg_config
    5. Install v8@3.15 by doing the following (these steps assume that vi/vim is the default editor):
      1. run HOMEBREW_NO_INSTALL_FROM_API=1 brew edit v8@3.15
      2. go to line 21 in the editor by typing :21 Note: the line being removed is disable! date: "2023-06-19", because: "depends on Python 2 to build"
      3. delete the line by pressing d twice
      4. save and quit by typing :x
      5. run HOMEBREW_NO_INSTALL_FROM_API=1 brew install v8@3.15
    6. Configure build opts for gem therubyracer:
      1. bundle config build.libv8 --with-system-v8
      2. bundle config build.therubyracer --with-v8-dir=$(brew --prefix v8@3.15)
    7. run ./scripts/dev_env_setup_step2.sh If you get a permission error while running gem install or bundle install, something went wrong with your rbenv install which needs to be fixed.
  3. If there are no errors messages, run bundle install to ensure all gems are installed

Running Caseflow

VERY IMPORTANT NOTE TWO: This is where you switch back to a standard (non-Rosetta) terminal

  1. Once your installation of all gems is complete, switch back to a standard MacOS terminal:
    1. open your ~/.zshrc file
    2. comment the line export PATH=~/homebrew/bin:$PATH
    3. uncomment the line export PATH=/opt/homebrew/bin:$PATH
    4. add the line export PATH=$HOME/.nodenv/shims:$HOME/.rbenv/shims:$HOME/.pyenv/shims:$PATH
    5. comment the lines eval "$({binary} init -)" for rbenv, pyenv, and nodenv if applicable
    6. if you added the line eval $(~/homebrew/bin/brew shellenv) after installing x86_64 homebrew, comment it out
  2. Open a new terminal and verify:
    1. that you are on arm64 by doing arch and checking the output
    2. that you are using arm64 brew by doing which brew and ensuring the output is /opt/homebrew/bin/brew
  3. Open caseflow in VSCode (optional), or navigate to the caseflow directory in your terminal and:
    1. brew install yarn
  4. Ensure you are on main branch and up to date by running
    1. git checkout main
    2. git fetch
    3. git pull origin main
  5. Start Vacols UTM VM (if not already running)
  6. run make up-m1 to create the docker containers and volumes
  7. run make reset to (re)create and seed the database; this takes a while (~45 minutes)
    1. if you get a database not found error, run bundle exec rake db:drop:primary db:create:primary db:schema:load:primary, and then run make reset again
  8. open a second terminal tab/window
  9. run make run-backend-m1 in one tab, and make run-frontend in the other
    1. the backend should launch on localhost:3000. go there in your browser to access Caseflow
    2. the frontend launches on localhost:3500; this is a server that enables hot-reloading of modules during development. going to this address in a browser will not work
    3. optionally, you can do make run-m1 to launch both at the same time with Foreman, but if one server errors the other can remain running and lead to headache trying to get them back up

To launch caseflow after a machine restart:

  1. Start the VACOLS VM via UTM; login is not required
  2. In terminal, run colima start to start the docker daemon
  3. In terminal, go to the caseflow directory and run:
    • make up-m1 to start the docker containers
  4. Open a second terminal window/tab, and from the caseflow directory run:
    • make run-backend-m1 in one tab
    • make run-frontend in the other tab

Note: It takes several minutes for the VACOLS VM to go through its startup and launch the Oracle DB service, and about a minute for the Postgres DB to initialize after running make up-m1.


Installation Workarounds

M3 Mac Gem Installation

Follow these steps if you are getting errors when running bundle install

  1. Create a new file script.py with the following contents:
     import subprocess
     import re
    
     #string = subprocess.check_output("gem query --local", shell=True)
     #string = re.findall("(?![^\\(]*\\))[A-Za-z-_]+", string.decode("utf-8"))
    
     with open("./gems.txt") as file:
         lines = [line.rstrip() for line in file]
    
     for i in lines:
     	print("bundle config build." + i + " --with-cflags=\"-Wno-error=incompatible-function-pointer-types\" --with-cppflags=\"-Wno-compound-token-split-by-macro\"")
     	output = subprocess.check_output("bundle config build." + i + " --with-cflags=\"-Wno-error=incompatible-function-pointer-types\" --with-cppflags=\"-Wno-compound-token-split-by-macro\"", shell=True)
    
    
  2. Create a new file gems.txt with the following contents:
     aasm
     actioncable
     actionmailbox
     actionmailer
     actionpack
     actiontext
     actionview
     activejob
     activemodel
     activerecord
     activerecord-import
     activerecord-oracle_enhanced-adapter
     activestorage
     activesupport
     acts_as_tree
     addressable
     akami
     amoeba
     anbt-sql-formatter
     ast
     aws-sdk
     aws-sdk-core
     aws-sdk-resources
     aws-sigv4
     backport
     benchmark-ips
     bgs
     bootsnap
     bourbon
     brakeman
     browser
     builder
     bullet
     bummr
     bundler-audit
     business_time
     byebug
     capybara
     capybara-screenshot
     caseflow
     choice
     claide
     claide-plugins
     cliver
     coderay
     colored2
     colorize
     concurrent-ruby
     connect_mpi
     connect_vbms
     connection_pool
     console_tree_renderer
     cork
     countries
     crack
     crass
     d3-rails
     danger
     database_cleaner
     date
     ddtrace
     debase
     debase-ruby_core_source
     derailed_benchmarks
     diff-lcs
     docile
     dogstatsd-ruby
     dotenv
     dotenv-rails
     dry-configurable
     dry-container
     dry-core
     dry-equalizer
     dry-inflector
     dry-initializer
     dry-logic
     dry-schema
     dry-types
     ecma-re-validator
     erubi
     execjs
     factory_bot
     factory_bot_rails
     faker
     faraday
     faraday-http-cache
     faraday_middleware
     fast_jsonapi
     fasterer
     ffi
     foreman
     formatador
     fuzzy_match
     get_process_mem
     git
     globalid
     govdelivery-tms
     guard
     guard-compat
     guard-rspec
     gyoku
     hana
     hashdiff
     heapy
     holidays
     httpclient
     httpi
     i18n
     i18n_data
     icalendar
     ice_cube
     immigrant
     jaro_winkler
     jmespath
     jquery-rails
     jshint
     json
     json_schemer
     kaminari
     kaminari-actionview
     kaminari-activerecord
     kaminari-core
     knapsack_pro
     kramdown
     kramdown-parser-gfm
     launchy
     libv8
     listen
     logstasher
     loofah
     lumberjack
     mail
     marcel
     maruku
     memory_profiler
     meta_request
     method_source
     mime-types
     mime-types-data
     mini_mime
     minitest
     moment_timezone-rails
     momentjs-rails
     msgpack
     multi_json
     multipart-post
     multiverse
     nap
     neat
     nenv
     net-imap
     net-pop
     net-protocol
     net-smtp
     newrelic_rpm
     nio4r
     no_proxy_fix
     nokogiri
     nori
     notiffany
     octokit
     open4
     paper_trail
     parallel
     paranoia
     parser
     pdf-forms
     pdfjs_viewer-rails
     pdfkit
     pg
     pluck_to_hash
     pry
     pry-byebug
     public_suffix
     puma
     racc
     rack
     rack-contrib
     rack-test
     rails
     rails-dom-testing
     rails-erd
     rails-html-sanitizer
     railties
     rainbow
     rake
     rb-fsevent
     rb-inotify
     rb-readline
     rchardet
     react_on_rails
     redis
     redis-actionpack
     redis-activesupport
     redis-classy
     redis-mutex
     redis-namespace
     redis-rack
     redis-rails
     redis-store
     ref
     regexp_parser
     request_store
     reverse_markdown
     rexml
     roo
     rspec
     rspec-core
     rspec-expectations
     rspec-github
     rspec-mocks
     rspec-rails
     rspec-retry
     rspec-support
     rspec_junit_formatter
     rubocop
     rubocop-performance
     rubocop-rails
     ruby-debug-ide
     ruby-graphviz
     ruby-oci8
     ruby-plsql
     ruby-prof
     ruby-progressbar
     ruby_dep
     ruby_parser
     rubyzip
     safe_shell
     safe_yaml
     sass
     sass-listen
     sass-rails
     savon
     sawyer
     scss_lint
     selenium-webdriver
     sentry-raven
     sexp_processor
     shellany
     shoryuken
     shoulda-matchers
     simplecov
     simplecov-html
     single_cov
     sixarm_ruby_unaccent
     sniffybara
     socksify
     solargraph
     sprockets
     sprockets-rails
     sql_tracker
     statsd-instrument
     stringex
     strong_migrations
     terminal-table
     test-prof
     therubyracer
     thor
     thread_safe
     tilt
     timecop
     timeout
     tty-tree
     tzinfo
     uglifier
     unicode-display_width
     unicode_utils
     uniform_notifier
     uri_template
     validates_email_format_of
     wasabi
     webdrivers
     webmock
     webrick
     websocket
     websocket-driver
     websocket-extensions
     xmldsig
     xmlenc
     xmlmapper
     xpath
     yard
     zeitwerk
     ziptz
    
  3. Move both files into the caseflow root folder
  4. In your Terminal, run python3 script.py
  5. Run bundle install again
  6. If any gems fail to install, manually install it by running gem install (e.g. gem install pg:1.1.4) and then run bundle install again to check for additional failures
  7. Once bundle install stops throwing errors, return to the last step for Script 2 (run ./scripts/dev_env_setup_step2.sh)

OpenSSL

When installing rbenv, nodenv, or pyenv, both openssl libraries should install as dependencies. Only follow the below instructions if you have problems with openssl@3 or openssl@1.1 not compiling.

  1. Download openssl@1.1 and openssl@3 from this link
  2. Open “Finder” and find the two folders under “Downloads”
  3. Extract the .tar.gz or .zip archives
  4. In each of the extracted folders:
    1. Navigate to the ~/homebrew/Cellar subfolder
    2. Copy the openssl folder to your local machine's ~/homebrew/Cellar folder
    3. If the folder Cellar in ~/homebrew does not exist, create it with mkdir ~/homebrew/Cellar
    • Note: moving these folders can be done using finder or a terminal
  5. Run command (from a rosetta terminal)
    1. brew link --force openssl@1.1
    2. If the one above doesn’t work run: brew link openssl@1.1 --force
    • Note: don't link openssl@3 unless you run into issues farther in the setup

Installing Ruby via Rbenv

If you are getting errors for rbenv being unable to find a usable version of openssl, run these commands prior to running the second dev setup script:

  1. brew install openssl@1.1
  2. export RUBY_CONFIGURE_OPTS="--with-openssl-dir=/usr/local/homebrew/Cellar/openssl@1.1"

Running Caseflow

The following steps are an alternative to step 7 of the Running Caseflow section in the event that you absolutely cannot get those commands to work:

  1. In caseflow, run
    • a. make down
      • i. Removes appeals-db, appeals-redis, and localstack docker containers
    • b. docker-compose down –v
      • i. Removes caseflow_postgresdata docker volume
    • c. make up-m1
      • i. Starts docker containers, volume, and network
    • d. make reset
      • i. Resets caseflow and ETL database schemas, seeds databases, and enables feature flags
      • **If make reset returns database not found error:
        • a. Run command bundle exec rake db:drop:primary
        • b. Download caseflow-db-backup.gz (not able to share this download via policy, ask in the slack channel)
        • c. Enter terminal, navigate to ~/Downloads
        • e. Run command
          • i. gzip -dck caseflow-db-backup.gz | docker exec -i appeals-db psql -U postgres
          • ii. (this command will link the caseflow_certification_database to docker)
        • f. Enter terminal, navigate to caseflow, and run
          • i. make up-m1
          • ii. make reset (this will take a while)

<< Back