Volume Data handling when backup fails

[droorda]

I don't know if this counts as a feature request.
I have a large volume that I am working on backing up (12TB). I have made several attempts to get the initial backup across a wan connection. From what I am seeing, the data that is sent every time is being added to a new vib file. When the job fails the space is not released and when the job retries it re-sends the same data. While I understand that the failed backup would not be usable for restoring data, it would be helpful if it was used by the dedupe code so that it did not have to be resent from the source. This would also allow for more efficient reties. As a less helpful solution could the vib have the unused space reclaimed?

[HannesK]

Hello,

From what I am seeing, the data that is sent every time is being added to a new vib file.

hmm, vib are incremental files. a VBK should be the first backup file.

The reference architecture is "local backup" for fast restores (well, also fast backups) and "backup copies" over WAN. For Backup copies, the WAN accelerator can be used.

If your WAN connection is fast enough for incremental backup (and restore...), then copy & paste, send disk via parcel service and backup mapping could be used.

Best regards,
Hannes