BackupEDGE 3.x - Quotas - Archive Retention - Lazy Reclamation

Quotas

During creation, each of the following BackupEDGE Resource types must be assigned a storage quota by the user.

• URL (FTP / FTPS Backups)
• S3CLOUD (Amazon and other S3 -compatible cloud storage sites)
• FSP (NFS Backups)
• FSP (CIFS Backups)
• FSP (Legacy Disk-to-Disk Backups)

BackupEDGE automatically assignes the quota as the actual device capacity on.

• SDRIVE (SharpDrive Backups)
• OPTICAL (DVD / BD-RE Backups)

BackupEDGE will not attempt to use more storage on the Resource than that assigned by the quota. An example mighty be 100GB.

Archive Expiration

By default, all archives have a retention time (or expiration time) of one week. This may be changed on a per-Schedule basis or even a per-backup basis. An archive may not be erased until the retention time is up. An archive past its retention time is called an Expired Archive.

Lazy Reclamation

Archives are not just automatically deleted on the same day of the next week, which was the default behavior in previous BackupEDGE releases. Lazy reclamation means that, once an archive has been created, two conditions must occur for it to be deleted...

• The retention time must be up, i.e. it must be an Expired Archive.
• Adding a segment to a new archive would cause the defined quota to be exceeded.

If both conditions are true, an Expired Archive will be deleted in its entirety. This process ensures that a maximum number of older archives are available.

If the Resource quota is reached and no Expired Archives are available for deletion, new backups will fail.

Beginnining with BackupEDGE 03.01.01 build 2, Lazy Reclamation may be disabled for SharpDrive, FSP, S3CLOUD and URL backups. If lazy reclmation is disabled, the next time a backup begins ALL expired archives on the Resource will be ereased.

Beginnining with BackupEDGE 03.01.02 build 2, Lazy Reclamation may be disabled for AWS  backups. If lazy reclmation is disabled, the next time a backup begins ALL expired archives on the Resource will be ereased.

Techical Detail

Backups are automatically split into segments. These are usually 1GB for URL and FSP-related Resources, and 50MB for S3CLOUD and AWS Resources. Adding the Re-startable flag to URL Resources also reduces the segment size to 50MB.

Every time BackupEDGE attempts to create a new segment on the Resource, it checks to make sure that the space it has used plus the space of the segment will be less than the quota. If not, a check for expired backups is performed and all segments from one of the expired backups is erased and the space returned to the quota pool.

This check is done at the beginning of every segment, and so may happen more than one time during a backup if necessary. This allows us to keep as many backups as possible around without hindering our ability to perform new ones. Who knows when you’ll have to restore a really old file that you thought was gone forever because it was outside a normal backup rotation. As it is done only when absolutely needed; we call it “lazy reclamation”.

Another major benefit is that if you are using a priced service like Amazon S3, where 100GB of storage is part of the base monthly service fee, you can set the quota to the space paid for and and be assured you’ll never exceed it. Of course, if your quota is too small to hold all your un-expired backups, segment creation, and ultimately the backup, will fail.

Last Updated - 2017/07/28