Author: paul.deluca

There are two methods of scheduling when CrashPlan runs:

“Always”

This is default setting. CrashPlan will always run, and will back up at the interval specified under Settings -> Backup tab -> Backup frequency and versioning.

The benefit to running CrashPlan “always” is that the backup will capture data throughout the day, therefore reducing the amount of data lost if a restore is necessary.

The downside to running CrashPlan “always” is that there is a potential for contention with the snapshot script, because both tools use different types of scheduling logic:

• The snapshot script is kicked off from crontab at regular intervals based on a static timeline (ex 12:00 AM, 3:00 AM, 6:00 AM, etc.).
• The CrashPlan backups are independently kicked off at regular intervals based on a dynamic timeline (ex 3 hours after the last backup to a particular destination completed).

Since the snapshots are kicked off at a regular interval via crontab, it’s possible that the snapshot could “refresh” itself (ie new snapshots are created) or “expire” (ie fill to capacity) during the CrashPlan backup process, thus invalidating the backup. The longer the CrashPlan backup(s) take, the higher the likelihood of the snapshot becoming invalid during the backup process. To prevent the snapshot script from running during an active CrashPlan backup, the snapshot script automatically stops the CrashPlan service during its operation. Upon completion of the snapshot creation process, the CrashPlan service is started, and the backup schedule is resumed.

An example configuration that has proven to work for many customers is the following:

• Set CrashPlan to run every 12 hours. The 12-hour interval between backups has shown to allow sufficient time for both local and remote backups to complete in a variety of configurations.
• Set the snapshot script to run at 12:00 AM and 12:00 PM. This ensures that a “point-in-time” representation of your data is available for restore at least twice daily.

“Between specified times”

There is an optional setting that can be configured under Settings -> General tab -> Crashplan will run: -> Between specified times.

The benefit to running CrashPlan “between specified times” is that it allows for the concrete definition of a window of time during which the backup should run. This is useful if there are bandwidth or performance constraints to running CrashPlan backups during business hours. The window should be made large enough to allow the completion of backups to each individual backup destination. The snapshot script should be scheduled to directly before the CrashPlan backup window, and the snapshot volume(s) must be large enough to remain active through the end of the backup window.

The downside to running CrashPlan “between specified times” is that the recovery points for data restoration are often reduced in conjunction with the available backup window. Despite this, the CashPlan agent still has many advantages over a traditional daily tape backup, including increased ease of administrator, archive reliability, security and the ability to backup to multiple destinations.

If you choose to modify the time during which CrashPlan operates, please ensure that the Eclipse snapshot backup script is scheduled via crontab to run prior to the CrashPlan backup window. For more information, see: How do filesystem snapshots work on Linux?