cloud-archive

A module to help with the task of archiving data tarballed, gzipped data to an object store (GS or S3 now). Includes retrieval

Environment

Endpoints and connection configuration are set via environment variables:

• OBJECT_STORE - the object store where tasks and grafs are stored and retrieved from
• FILE_NAME_FORMAT - uses templating to specify a pattern for creating tarball names. Tokens are escaped with {{ }}.
  • All dates and times for tokens are UTC
    • dateTime - provides date and time: mm_dd_yy_HH_MM_SS
    • date - provides date: mm_dd_yy
    • time - provides time: HH_MM_SS
• FILE_PATTERNS - a comma delimited list of glob patterns to tar when creating a backup. defaults to **/*
• BASE_PATH - defaults to the processes' path
• DATA_PATH - a subfolder off the base path, defaults to archive
• FILE_NAME - a default archive name to use when restoring rather than finding the most recent. Useful in cases where you want to restore from a known archive.

AWS:

• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY

GS:

• GS_PROJECT_ID
• GS_USER_ID
• GS_USER_KEY

Archive limitations:

• DISCARD_AFTER - how many days to keep archives for before removing them
• COLDLINE_AFTER - how many days before a file is moved to "cold storage" this is primarily offerred as a way to move unliked files into a cheaper tier of slower/low priority storage.

API

backupFrom([glob])

Backs up files from dataPath and can selectively filter for files using the optional glob argument.

Returns a promise that will resolve on success with the filename or reject with an error on failure.

restoreTo([fileName])

Download the latest (or optionally, a specific filename) to dataPath.

Returns a promise that will resolve on success with the full path and file listing or a reject with an error on failure.

On success, the resulting data structure looks like:

{
  path: '/full/path/to/files',
  files: [
    'file1.txt',
    'file2.txt',
    'file3.txt',
    'file4.txt'
  ]
}

TODO

• GS_USER_KEY expects PEM, not JSON key