This article applies to CrashPlan PROe version 3.
Other available versions:
This tutorial explains how to expedite backing up a large amount of data through a process known as "seeding."
If you have a large amount of data to back up to an offsite private cloud destination, backing up over the Internet can be very slow and can potentially saturate your connection. You can avoid this by using a process called seeding. When backups are seeded, devices back up to a local destination and then the archives are copied or transported to a store point at the offsite private cloud destination. Only new files and incremental changes, which are typically small, are sent across the Internet going forward and subsequent backups take much less time.
Enterprise server seed directory
On every enterprise server, there is a service that checks each store point every 5 minutes for backup archives placed in a special seeding directory called CrashPlanArchive_SEED. The CrashPlanArchive_SEED directory is paired with a store point. The seeding directory resides at the same level as the store point's directory:
- Linux: /var/opt/proserver/backupArchives/CrashPlanArchive_SEED
Applies to enterprise servers installed as root on Ubuntu
- Windows: C:\ProgramData\PROServer\backupArchives\CrashPlanArchive_SEED
- OS X: /Library/Application Support/CrashPlan/PROServer/backupArchives/CrashPlanArchive_SEED
- Solaris: /var/opt/proserver/backupArchives/CrashPlanArchive_SEED
A backup archive in the CrashPlanArchive_SEED directory is moved onto the associated store point once the backup archive remains the same size two consecutive seed checks. This is to ensure that the move or copy into the CrashPlanArchive_SEED directory has completed before a backup archive is moved to the store point. It may take up to 10 minutes for the server's seed service to begin moving an archive from the CrashPlanArchive_SEED directory to the associated store point.
Before you begin
Before you begin the seeding process, verify the following:
- There is enough space for all backup archives you are seeding on the destination server
- The device or devices you are seeding are active
- The destination you are seeding to is offered to the device or devices you are seeding
CrashPlan PROe users with the admin.system.command permission are allowed to configure seeding settings. The following default roles contain this permission:
- SYSADMIN role
- Server Administrator role
- Seeding can only be done in installed environments. Managed Private Cloud (managed appliance) environments cannot seed.
- The seeding process described in this article is for CrashPlan PROe version 3.5 or later.
For CrashPlan PROe version 3.4.x, please see the 3.4.x seeding instructions.
- Seeding is intended for a device's initial backup only. The seeding process completely overwrites any data previously backed up to the offsite Private Cloud destination for any devices you are seeding.
- If data balancing is enabled on the store point the archive is being seeded to, the data balancer may move a seeded archive immediately upon seeding completion.
- Once a backup archive is placed in the CrashPlanArchive_SEED directory, the archive is moved to the associated store point. This is a move operation and not a copy. Take care to ensure you have another copy of the device's backup archive.
- When copying existing backup archives, always verify the copy operation using checksums or another verification method.
- Archive maintenance occurs on an archive immediately after the completion of the copy operation. You may be able to speed up the maintenance process by increasing the maintenance rate for system jobs at Settings > Server > General > Archive maintenance queue > Action Menu > Adjust Maintenance Rates.
- Choose a server within the private cloud destination where you'd like the backup archives seeded.
- On the server you've selected, sign into the administration console.
- Specify an administrator email address to receive seeding success/failure messages:
- Double-click the CrashPlan PROe logo in the upper-left corner of the administration console to open the PROe Server CLI.
- Set the
c42.seed.emailproperty by issuing the following command:
prop.set c42.seed.email email@example.com
- Create the seed or seeds:
Back up the device or devices you are seeding to some local destination. This can be a server destination, a PROe Client within the same user account, or a local folder.
- Copy the seed archive(s) created at the local destination into a staging directory on the same volume as the store point you are seeding.
- Copy the seed archive(s) into place.
- On all store points, there is a directory called CrashPlanArchive_SEED at the same level as the store point directory.
- The user that runs the PROe Server service must have permission to write to the CrashPlanArchive_SEED directory.
- Copy the seeded archive(s) into the CrashPlanArchive_SEED directory.
You may need to wait up to 10 minutes before the seeding job begins moving a seeded backup archive from the CrashPlanArchive_SEED directory to the associated store point.
If there is a problem during the seeding process, the server where you are seeding sends an email to the address configured in that server's c42.seed.email system property. The affected seed archive is renamed GUID.SEED_FAILED.timestamp and resides within the CrashPlanArchive_SEED directory. E.g.
There are several possible error messages in cases where the seed service cannot complete the seeding process.
|Computer not found||The device does not exist within the master server's database. The device must be active within your CrashPlan PROe environment in order to seed. Backup archives created under another CrashPlan PROe environment (with a different master server) cannot be seeded.|
|Destination not offered||The device being seeded is not allowed to back up to the destination that is being seeded. Enable backup to the destination via the organization level or global level destination settings.|
|Unknown destination error||Problem gathering information about the destination where the seeded archive has been placed.|
|Unreadable file||Indicates either a permissions issue or corruption in the seed archive.|
|System error||Catch-all for errors that do not fall into the above categories. System errors also include a stack trace for additional diagnostic information.|
If your environment is running version 3.4.x, please follow these instructions to seed the backup to your own private cloud.
- Create the seed:
Back up the device you are seeding to some local destination. This can be a server Destination, a PROe Client within the same user account, or a local folder.
- Assign the device to a store point on the remote destination:
Start backup from the device to the desired remote destination. This assigns the device to a store point and creates a backup archive.
- Place the seed in the proper location on the remote destination:
- Go to the store point where the device was assigned
- Disable balancing on the store point
- Pause backup for the device you are seeding
- Replace the archive created in Step 2 with the seed
- Resume backup to the remote destination.
The device synchronizes and backup begins where it left off during the creation of the seed once syncing is complete.
- Re-enable balancing on the remote destination's store point