Enhance Backups Method

[ferricoxide]

Is your feature request related to a problem? Please describe.

Per #22, backup job, as currently architected, can unpredictably run away. This has become a more frequent condition as more users execute multi-branch pipelines (which run on the master and cause target-files to appear and disappear during backup runs). Likely, it's either the overall business of the master or the business of the filesystem being backed up that's causing the runaways.

Describe the solution you'd like

Replace the simple, one-line, "dumb" tar-direct-to-S3 backup job with something more intelligent. Likely, this would include:

• Changing ${JENKINS_HOME} to be hosted on an LVM (vice its current, "bare" /dev/sdX device)
• Creating an actual backup script to be called from crond that:
  1. Adding locking-logic to script to ensure only one job at a time runs (and that any prior failed runs are cleaned up)
  2. Creates an LVM snapshot of ${JENKINS_HOME}
  3. Backs up the LVM snapshot of ${JENKINS_HOME}
  4. Destroys the LVM snapshot once backup completes
  5. Removes lockfile