#1 2017-12-29 18:56:25

admin
Administrator
Registered: 2017-04-21
Posts: 1,363

OneDiff prog: mismatch in used size comparison with thick disks

When you use (c) OneDiff to backup a VM containing thick disks (lazy zeroed or eager zeroed) to a local datastore (VMFS), the quick Trivial Check performed after the backup shows a used size mismatch between the original disk and the just backed up one.

This is due to vmkfstools always generating a thin disk on the backup target (to take advantage of available space), so when the original disk is queried for used room, it always returns its full size, whereas the backed up disk only returns the used size in the thin sparse file.

Workaround:

While we release a fix for these cases, you can use the --certify-backup=yes argument, which will change the Trivial Check by a full checksum comparison on the disks.

The size mismatch error can be safely ignored if no other errors are present, in fact it would be like OneDiff'in to an NFS datastore with a different FS, whith the only difference that in that case XSIBackup detects the file system not being VMFS and just skips the used size check.

Offline

Board footer