As subject... the speed-up it gives is Massive, so I would like to use it, but I also want reliability.
Another imaging product seems to do a chkdsk before the backup when using a similar fast backup - I assume to minimise risk.
- Does IFW do chkdsk before backup ?
- What are the potential problem's of using metadata backups ?
- Any other thing's I should know from anyone who's researched/used it ?
Opinions wanted on reliability of Metadata Hash backup feature ?
Re: Opinions wanted on reliability of Metadata Hash backup feature ?
The warning about using metadata are often because metadata might be incorrect, resulting in a bad backup, but...
- Ultimately, the OS is using the metadata itself when working with the file-system/disk, albeit with extra sanity check's that the Win32 API's do.
- Given that, aside from the extra sinty check's the API's do, is there any other reason not to use metadata option (particularly if chkdsk is done before-hand) ?
Trying to properly understand why it's not default feature, rather than optional. As mentioned - for backup's, reliability is #1, critical feature, above all else.
- Ultimately, the OS is using the metadata itself when working with the file-system/disk, albeit with extra sanity check's that the Win32 API's do.
- Given that, aside from the extra sinty check's the API's do, is there any other reason not to use metadata option (particularly if chkdsk is done before-hand) ?
Trying to properly understand why it's not default feature, rather than optional. As mentioned - for backup's, reliability is #1, critical feature, above all else.
-
TeraByte Support
- Posts: 3625
- Joined: Thu May 05, 2011 10:37 pm
Re: Opinions wanted on reliability of Metadata Hash backup feature ?
There is a topic in here that explains why you must use vss and not mix with phylock if using it.
Re: Opinions wanted on reliability of Metadata Hash backup feature ?
I've read that and also believe IFW uses VSS by default (when Phylock was default snapshot provider in the past).
I'm still wondering why metadata feature isn't used by default, along with answers to other question's ?
I'm still wondering why metadata feature isn't used by default, along with answers to other question's ?
-
pimjoosten
- Posts: 54
- Joined: Sat Jul 13, 2019 3:37 pm
- Location: The Netherlands
Re: Opinions wanted on reliability of Metadata Hash backup feature ?
I always use the metadata hash option when making incrementals and differentials and almost always use it when restoring backups. I have never seen any adverse effect. One of the main points I took from the thread that was mentioned is that the risk associated with the metadata hash option lies with a restore operation using the metadata, as Brian K concludes in this post: viewtopic.php?p=13872#p13872. "The metadata restore option should only be used for undoing changes and not for recovering from file system corruption", as the latest manual (2023-07-08) states on p. 107. In the latter case I always do a full restore.
About the chkdsk performed by the other product: that is not really a chkdsk but it checks (quote) "that every single object (file) in the file system has all data (clusters) mapped as 'In Use' by the NTFS meta data file $BITMAP. If there are any other file system integrity errors then these will persist in the image file, but crucially there will be no clusters (data) excluded from the Image. In this case any file system integrity errors will be repairable. If the $BITMAP file has 'used' clusters marked as 'Free' then this can be catastrophic for the file system during normal operation, and will also mean that data could be excluded from the Image. This is what the verification checks for." Such a check is relevant for all backups that only backup the sectors in use (instead of a raw backup), not just the ones that are made quickly using metadata. It would indeed be very interesting to know if and how IFWDLU checks what I described above before making a backup.
About the chkdsk performed by the other product: that is not really a chkdsk but it checks (quote) "that every single object (file) in the file system has all data (clusters) mapped as 'In Use' by the NTFS meta data file $BITMAP. If there are any other file system integrity errors then these will persist in the image file, but crucially there will be no clusters (data) excluded from the Image. In this case any file system integrity errors will be repairable. If the $BITMAP file has 'used' clusters marked as 'Free' then this can be catastrophic for the file system during normal operation, and will also mean that data could be excluded from the Image. This is what the verification checks for." Such a check is relevant for all backups that only backup the sectors in use (instead of a raw backup), not just the ones that are made quickly using metadata. It would indeed be very interesting to know if and how IFWDLU checks what I described above before making a backup.