The support forum

Any update on the verification feature?

cyrux :

Jan 26, 2018

Hey, any update on the verification feature? It seems to have been on the horizon for a veery long while, and seems like a really core feature that should have been a priority to implement, even if in an imperfect form?

Froggie :

Jan 27, 2018

No, you don't implement PRIORITY items in an imperfect form  :)

cyrux :

Jan 27, 2018

Imperfect doesn't mean broken or useless.

At some point the dev mentioned his ideal solution would to store hashes after the first run to save time on subsequent verifications. A lesser but perfectly adequate solution would just be recalculating hashes each time you run verification, and would be much simpler to implement. Such a function would probably be greatly appreciated by the majority of users, since honestly running verification once or twice a year is going to be more than enough.

In my opinion it isn't worth the delay (standing well over a year now) to implement the more complex solution, and as this feature is so important it would be nice to hear where it is in the pipeline.

Alex Pankratov :

Jan 27, 2018

Quite a few features have been on the todo list for a while. For the sake of providing a context - in addition to the backup verification there are also parallel copying, licensing changes, UI support for job grouping and SMART monitoring for local storage devices. Additionally, there's a handful of smaller items to add UI presence for already supported features, like scheduling queues, quiet times and media ejection.

The hold up is primarily due to me wanting to move us to the feature-based licensing model before we add any more large features. Once the licensing changes are in (Release 79), the rest will follow and it should follow quickly.

Parallel copying is nearly done. In fact, it's not just _copying_ anymore. We now have a new revision of the backup engine that parallelizes *all* backup operations including prep- and post-stages, e.g. extraction of file IDs, mass-deletion of files and syncing folder timestamps.

Furthermore, Diskovery's "storage stack comprehension" library got folded in Bvckup 2, so it now knows exact types of storage devices and it can reconfigure the copying pipeline to match... In other words, this was a very big chunk of work, not visible publicly, but spanning several months from August last year. This too will be R79.

Backup verification will follow next. Not that same release, but shortly after. The good news is that the new backup engine is easily extensible, so equipping it with verification logic is quite simple. The main hurdle here actually is to find the real estate for verification settings in the UI... meaning that it may require some UI rework which tends to be a time-consuming task.

PS. With regards to checksumming - the point of keeping checksums is to allow telling apart bitrot at the source from the bitrot at the backup. If you are looking at a checksum mismatch, you'd better be damn sure that it's the backup that spoiled and not the source.

However this is a very minor technical nuance and the need to do that is not the reason for the delay.

New topic

Made by Pipemetrics in Switzerland

Miscellanea Press resources
On robocopy

Legal Terms