How can we help?

We think these articles could help:

    See More
    Home > CrashPlan > Latest > Backup > How Backup Works

    How Backup Works

    Applies to:
    • CrashPlan for Home
    • CrashPlan PRO
    • CrashPlan PROe

    Overview

    Have you ever wondered what goes on with CrashPlan behind the scenes? In this article, we'll take a closer look at exactly what happens when CrashPlan is backing up.

    Backup Basics

    Let's start with an example scenario where CrashPlan is using its default settings and is backing up your user home directory.

    Home Directory Selected For Backup

    CrashPlan constantly watches for new and changed files within your home directory with what we call the real-time file watcher. It adds new and changed files to a TO DO list.

    Let's say that you created a document called “Letter to Grandma.doc”. The real-time file watcher sees that you've created this document and adds it to the TO DO list for backup.

    One File On Backup To-Do List

    This is what happens when CrashPlan starts backing up Letter to Grandma.doc:

    1. Backup begins with a process called data de-duplication. CrashPlan analyzes a small piece of the file (a block), and checks to see if that block was backed up previously.
    2. If CrashPlan determines that it has already backed up this block, CrashPlan moves on and analyzes the next block.
    3. If the block has not yet been backed up, CrashPlan:
      1. Compresses the block to save storage space
      2. Encrypts the block to secure the data
      3. Sends the block to the backup destination

    Backup Process: New File

    Data is securely encrypted throughout this process.

    The process repeats for the next block within the file until CrashPlan has analyzed and backed up the entire file. In this way, only unique information is backed up, which saves bandwidth and storage, and makes restoring faster

    Data De-Duplication 
    Data de-duplication occurs on each computer. If you have the same file on two different computers, the file will be backed up twice—once for each computer.

    New Files And File Changes

    As you're working on your letter and making changes, CrashPlan's real-time file watcher sees that the file has changed and CrashPlan puts the file back into the TO DO list. If your letter is 1 MB, 1 MB is added to the TO DO list. Only the changes are actually sent to the destination, however, not the entire file. The changes are backed up while you work, creating a new version of “Letter to Grandma.doc”.

    In this example, you've added a paragraph (highlighted in red):

    Backup Process: File Changes

    1. CrashPlan's data de-duplication scans the file looking for new blocks of data.
    2. The new (red) data blocks are:
      1. Compressed to save space
      2. Encrypted for security
      3. Transmitted to the backup destination for storage
    Backing Up New File Versions  
    By default, a new version of the file is backed up every 15 minutes. This interval is controlled by the New version setting. We recommend keeping the 15-minute default in most cases.

    Detecting Changes - The Details

    There are actually two ways that CrashPlan learns about new files or changes to your existing files:

    1. Real-time file watcher
    2. File system scan

    With two methods of identifying file changes, your files are doubly protected—CrashPlan checks for changes twice to make sure your files are backed up.

    The real-time file watcher works directly with the tools built into your computer's operating system, which means it is fairly lightweight and can easily work in the background without you noticing.

    The scan requires a bit more resources than the real-time file watcher, so to minimize possible impact on your computer, the scan runs at 3 am every day by default, when most people are less likely to be working at the computer.

    Detecting Changes On OS X
    On OS X, CrashPlan detects that you've deleted a file or files during the scheduled scan. On other operating systems, deleted files are detected in real-time.

    Prioritizing Files For Backup

    Of course, you probably have more than one file on your computer that you'd like backed up. CrashPlan can't back up all files simultaneously, so how does it decide what to work on first? We designed CrashPlan to back up the newest and most recently changed files first. This ensures that the most recent versions of your files—what you're working on right now—are backed up as soon as possible. The priority order looks like this:

    1. Newer, smaller files
    2. Newer, larger files
    3. Older, smaller files
    4. Older, larger files

    Whenever a file is added or modified, CrashPlan adds it to its backup to-do list. Changes are added to the backup based on the frequency at which CrashPlan backs up new versions, which by default is every 15 minutes. You can change this default setting under Settings > Backup).

    Backing Up Very Large Files
    If you have very large files that change frequently (such as multiple-GB virtual machine disks) and it seems like backup never completes, try creating a backup set for those large files with a longer New version interval. This gives CrashPlan more time to back up other files before it needs to back up the changes within the very large file or files.

    Prioritizing Backup To Multiple Destinations

    We always recommend that you back up to multiple backup destinations for fastest backup and restore and for best protection, but exactly how does multi-destination backup work?

    CrashPlan prioritizes backup activity to ensure all your selected files are completely backed up at one destination before starting backup to another, redundant destination. To accomplish this, CrashPlan backs up to destinations which it determines should complete fastest:

    1. Local folders
    2. Computers on the same network (LAN)
    3. Online or to computers across the Internet (WAN)

    For example, if you are backing up to a local folder and to a friend's computer, CrashPlan completes backup to the local folder before backing up over the Internet to your friend's computer. Once backup to a destination completes (or if that destination becomes unavailable for any reason), CrashPlan backs up to the next destination.

    Restoring A File 
    You don't have to wait for the entire backup to complete to restore a file. As soon as a file is backed up and appears on the Restore tab, it is available for you to restore.

    Backup Sets

    If you choose to enable Backup Sets, you can specify backup priority. The goal is still to back up all your files to at least one destination first. Then, CrashPlan works on redundancy to back up your files to additional destinations. As long as one destination in the set is complete, CrashPlan moves on to back up other, less complete sets. When you have Backup Sets enabled, CrashPlan follows these rules:

    1. Backup Set priority
    2. Destination's connection type (local folder, then LAN/local network, then WAN/across the Internet)
    3. Percent complete to that Destination
    Multiple Backup Sets With A Single Destination
    When multiple backup sets back up to a single destination, there are some special considerations for file exclusions and frequency and version settings.

    Multiple Computers And Backup Data Storage

    The backup process always works the same, whether you have one computer or ten computers in your account.

    CrashPlan treats each computer within your account separately and each computer's backup is stored separately at each destination. Each source computer stores its backup in its own folder. The folder name is the source computer's ID.

    In this example, John and Michelle are backing up to Chris. This is what it looks like on Chris' computer:

    Friend Computer IDs And Backup Folders

    Storing each computer's backup in its own folder enables seeding, which is the process of backing up locally and then physically transporting the backup drive to an offsite destination, such as a friend across town or to our data centers. Seeding is a great way to kick-start a large backup to an offsite destination and saves on bandwidth because the backup does not go over the Internet.

    Is My Backup Starting Over?

    Occasionally, CrashPlan's data de-duplication needs to re-scan your files to see what's already been backed up. When this happens, it looks like CrashPlan is backing up all your files from the beginning, but it is actually reviewing each block to see what's been backed up already. How can you tell that this is the behavior you're seeing?

    • Progress is much, much faster than a full initial backup because blocks that have already been backed up are not re-sent.
    • All your files are available for restore during this process.
    • Verify that the amount of data being stored is a number that is consistent with your file selection size and previous backup completion. From Backup, click the name of the destination to view stats for the chosen destination. Under Space used, verify that this number is an amount of data reasonable for your file selection size and previous backup completion.

    CrashPlan's Cache
    CrashPlan's cache includes information on de-duplicated data. You'll experience the above behavior if CrashPlan needs to rebuild its cache for any reason. This is something that happens on occasion under normal use.
    You must to post a comment.
    Last Modified
    16:47, 19 Nov 2014

    Tags