Skip to content

Latest commit

 

History

History
303 lines (239 loc) · 6.2 KB

backup-your-github-repos.md

File metadata and controls

303 lines (239 loc) · 6.2 KB

How-To Back Up Your GitHub Repos - A Step-By-Step Guide

Why would you want to back up your github repos? Isn't github itself a backup (even with an arctic vault option - see How the cold storage will last 1000 years)?

A personal anecdote - in the morning (on February 15th) I was preparing a tech talk for the FreeCodeCamp Vienna (on programming pixel art in, yes, Ruby - Big in Japan! - and JavaScript) and heading out to Vienna in the afternoon for the live in-person talk.

When I checked-up the next day in the morning to post a link to the github repo it was gone (and a dozen more) - all taken down (with a 24-hour notice) by an anonymous crypto bro ("More Punks") filling out an online form (claiming the sole copyright to 24×24px image(s) in .png or even ascii art).

While I had local (up-to-date) working copies for recovery for some taken down repos that I was working on in the last days - for others the story is more complicated.

For more on the (ongoing) crypto punk's not dead under attack take down story see the (Crypto) Punk's Not Dead org.

Anyways, let's use ruby (and git) to back up your repos using git (& github) gems from the ruby code commons (cocos) family:

  • gitti-backup gem - (yet) another (lite) git backup command line script and
  • hubba gem - (yet) another (lite) GitHub HTTP API client / library.

Step 1: Prepare A List Of Repos (By User / Org) To Backup

If you have only a couple of repos - you can prepare the repos list (in the yaml format) by hand or let's automate the chore / task querying the github api (via the hubba gem) for all repos.

Example - prepare.rb:

require 'cocos'
require 'hubba'

## get all repos for the user geraldb  (exclude all orgs)
h = Hubba.reposet( 'geraldb', orgs: false )
pp h

write_text( './repos.yml', h.to_yaml )   ## save as yaml

resulting in:

geraldb (4):
- geraldb
- geraldb.github.io
- sandbox
- talks

Let's add another user (for example, an alter-ego automation bot account) or your business partner or friend etc.

## get all repos for the user geraldb and yorobot (exclude all orgs)

h = Hubba.reposet( 'geraldb', 'yorobot', orgs: false )
pp h

write_text( './repos.yml', h.to_yaml )

resulting in:

geraldb (4):
- geraldb
- geraldb.github.io
- sandbox
- talks
yorobot (18):
- auto
- backup
- beer.db
- cache.csv
- cache.github
- football.csv
- football.db
- football.db.rsssf
- football.json
- logs
- planetjekyll
- rubyconf
- sport.db.more
- stage
- tipp3
- workflow
- world.db
- yorobot

Now if you have many of your repos grouped by org(anization)s and NOT under your personal account drop the extra orgs: false keyword parameter and the hubba machinery will (automagically) query for all orgs and than query for all orgs one-by-one of all repos.

h = Hubba.reposet( 'geraldb', 'yorobot' )
pp h
write_text( './repos.yml', h.to_yaml )

resulting in:

geraldb (4):
- geraldb
- geraldb.github.io
- sandbox
- talks
yorobot (18):
- auto
- backup
- beer.db
- cache.csv
- cache.github
- football.csv
- football.db
- football.db.rsssf
- football.json
- logs
- planetjekyll
- rubyconf
- sport.db.more
- stage
- tipp3
- workflow
- world.db
- yorobot
austriacodes (5):
- austria.txt
- austriacodes.github.io
- awesome-austria
- showcase
- vienna.html
beerbook (3):
- beerbook.github.io
- calendar
- maps
beercsv (7):
- be-belgium
- build
- ca-canada
- de-deutschland
- statistics
- us-united-states
- world
beerkit (9):
- beer.db
- beer.db.admin
- beer.db.day
- beer.db.labels
- beer.db.labels.ruby
- beer.db.service
- beer.db.starter
- beer.js
- beerbook
bibtxt (1):
- bibtxt.github.io
bigkorupto (3):
- awesome-nocode
- mammad-kabiri-uniqa
- sources
bitsblocks (11):
- bitcoin-maximalist
- bitcoin-whitepaper
- bitsblocks.github.io
- colored-coins-whitepaper
- crypto-bubbles
- crypto-facts
- ethereum
- ethereum-whitepaper
- get-rich-quick-bible
- islandcoin-whitepaper
- mastercoin-whitepaper
# and on and on ...
yukimotopress (15):
- auto
- blockchains
- examples
- fizzbuzz
- gem-dev
- gem-tasks
- http
- langs
- micro
- practices
- practicing
- sinatra-intro
- smalldata
- start
- yukimotopress.github.io

Step 2: Back Up Your Repos Via Git Mirror ("Bare Bone")

Once you have prepared your repos list - prepare the backup script. Example - backup.rb:

require 'gitti/backup'

## read-in the repo list
reposet = GitRepoSet.read( './repos.yml' )

## save all backups in /backup root  or wherever you want
backup = GitBackup.new( '/backup' )

## run the backups via git clone --mirror
##   if repo is new / first-time "bare bone" clone,
##   otherwise sync via  git remote update
backup.backup( reposet )

puts 'bye'

resulting in:

[1/551] geraldb@geraldb...
---> (shell run) >git clone --mirror https://github.com/geraldb/geraldb<... OK
Cloning into bare repository 'geraldb.git'...
[2/551] [email protected]...
---> (shell run) >git clone --mirror https://github.com/geraldb/geraldb.github.io<... OK
Cloning into bare repository 'geraldb.github.io.git'...
[3/551] geraldb@sandbox...
---> (shell run) >git clone --mirror https://github.com/geraldb/sandbox<... OK
...

and your directory tree like:

/backup
  /geraldb
     /geraldb.git
     /geraldb.github.io.git
     /sandbox.git
     ...
  ...

Note: Yes, you can (re)run the backup script anytime to update / sync the (bare bone) git (backup) repos.

resulting in:

[1/551] geraldb@geraldb...
---> (shell run) >git remote update<... OK
Fetching origin
[2/551] [email protected]...
---> (shell run) >git remote update<... OK
Fetching origin
[3/551] geraldb@sandbox...
---> (shell run) >git remote update<... OK
Fetching origin
...

for example, if there no changes and your backups are up-to-date.

That's it.