git_transactor

Overview

This code performs a series of git add and git rm transactions on a local repository, and then pushes the updates to a remote repository.

Preconditions

• A git client is installed on the system.
• The user running the script has write privileges on both the local and remote repositories.
• The user running the script has read privileges on files being added to the local repository.

Status

in production

Available Functionality

rake is used to execute git_transactor tasks.

rake git_transactor:process_and_push   # Process entries in queue and push to remote
rake git_transactor:pull               # Pull from remote repository to local repository
rake git_transactor:pull_process_push  # Pull from remote repository, process queue, and push back to remote repository
rake git_transactor:push               # Push from local repository to remote repository
rake git_transactor:setup:queue        # Initialize queue structure in QUEUE_ROOT directory

The following environment variables must be set when running a git_transactor rake task:

GT_LOCAL_REPO       # the path to the local EAD repository
GT_SOURCE_PATH      # the path to the directory from which to copy EAD files
GT_QUEUE_ROOT       # the path to the Git Transactor queue directory structure created via
                    #   rake git_transactor:setup_queue task
GT_REMOTE_REPO_URL  # the remote repository URL

e.g.,
rake git_transactor:pull_process_push LOCAL_REPO='./super_duper_files' \
  SOURCE_PATH='/path/to/source/files' QUEUE_ROOT='/path/to/queue/root' \
  REMOTE_REPO_URL='git@github.com:example/super_duper_files.git'

pull_process_push execution overview

1. as the name implies, the task first executes a pull to synchronize the local repo with the remote repository
2. then, a process_queue operation is executed. process_queue operation does the following:

• checks for "queue entry files" in QUEUE_ROOT/queue directory
• if the QUEUE_ROOT/queue directory is empty the script simply terminates with an exit status of 0
• if the QUEUE_ROOT/queue directory is not empty, the script processes each queue entry file in the order of the queue entry filenames, from lowest to highest.
• if a request is processed successfully, the queue entry file is moved to the QUEUE_ROOT/passed/ subdirectory
• if a request is processed unsuccessfully, the queue entry file is moved to the QUEUE_ROOT/failed/ subdirectory.
• after all queue entry files, the changes are committed to the local repository

3. the task next executes a push operation to synchronize the remote repository with the local repository

Queue Entry Files

Each queue entry file is named per the following template:

<YYYYMMDDTHHMMSSNNNNNNNNN.csv>, where:
  * YYYY = current year
  * MM   = current month,  zero-padded, range: 01-12
  * DD   = current day,    zero-padded  range: 01-31
  * T    = literal character T used as date/time delimiter
  * HH   = current hour,   zero-padded  range: 00-23 (note 24-hour format)
  * MM   = current minute, zero-padded, range: 00-59
  * SS   = current second, zero-padded, range: 00-59
  * NNNNNNNNN = current nanosecond, zero-padded, range: 000000000-999999999
  e.g., 20150227T113018273091611.csv

A queue-entry file contains a single line of comma-separated values that conform to the following template:

<command>,<absolute path to file being managed>
<command> must be either add or rm