Double destination for backup

First of all, congratulations on your software.

I am a very careful person to backup copies of my work - in the past I have lost some work done - and I will find it very useful if there was the possibility to set a double destination for the backups.

If possible, I would set up a first location for backups on my computer’s hard disk and a second location for backups on a cloud drive (for example Dropbox).
In this way, if I lose my job due to hard disk failure, I have the copy in the cloud; if the copy to the cloud fails, I have the work saved on the hard disk; if I break the hard disk and the cloud copy failed… it wasn’t a good job :wink:

You could set a backup to the DropBox (or other cloud) folder on your hard drive which will be synced with Dropbox giving a local and cloud copy. The only issue a creeping issue with a hard drive could see damaged backups copied from the local folder to Dropbox.

With the reliability of drives, especially SSDs it’s unlikely to be an issue.

I use the above approach, plus Time Machine, perhaps one of the best backup apps on any platform. There are likely similar for Win.

I throw in a bootable backup from time to time to be sure to be sure to be sure…. (apologies to my Irish friends)

If use dropbox by definition the backup is in the cloud and on your computer on dropbox. I also use carbonite as 3 rd backup and this automatically backs up everything on my computer (no size limit and have over 700 gigs backed up including pictures, music, manuals, etc) for 60 dollars a year for one computer and i back up everything important in the house to one computer to take advantage. I look at it as digital insurance. I don’t want to lose my scrivener work, nor my family pictures or 200 gigs of music and worked well when had hard drive failure a few years ago. We pay car and house insurance, I consider carbonite data insurance and covers scrivener as well. (I get nothing from carbonite)

1 Like

Just to be clear, Dropbox in the default configuration is not a backup. When you make a change to the on-disk copy under your Dropbox folder, the version on Dropbox’s servers in the cloud will be updated as soon as the sync engine runs.

Just having another copy does not make it a backup unless it is isolated from changes that occur to main copy. Carbonite is an excellent backup program!

1 Like

It is if it is a backup you are saving to the local Dropbox folder, which is what I and I believe others were referring to above. The Backup won’t get written to after saving, therefore neither it or the Dropbox copy get modified. (Unless you delete the local copy of course) Your statement would of course be true of a SAVE to Dropbox.

I have an issue with paying a subscription for backups when there are readily available low-cost external drives, NAS etc to do it locally and keep control. The Carbonite sub would allow me to add 1-2TB per year to my local backups.

As I’ve said before, Time Machine is an amazing app and free on macOS. I haven’t tried any of the Win ‘equivalents’ but this is just one of a number of listings from a quick DuckDuckGo

1 Like

The distinction I am making — and that we on this forum seem to have conversations around time and again — is that if something changes that Dropbox-synced folder (whether the live project or a backup copy) then Dropbox sync will happily propagate that change and destroy those copies. Dropbox core does not make point-in-time copies which is what keeps it from being a backup.

Now, if you have some sort of versioning enabled (so that each change is preserved separately instead of completely overwriting the base file) then that’s a different story. At work I just had a critical file get destroyed (and that destruction propagated to the cloud) because OneDrive for Business happily synced the replacement of the source file with a zero-length “update.”

Luckily my employer enables version history on all personal ODfB archives, so I was able to retrieve and download the previous version and restore that file to what it was before the 0-byte “update” occurred. However, for a Scrivener project — which does NOT look like a single file to cloud services, regardless of how it appears in MacOS — this would still be an extremely painful process to recover from unless you were recovering a single-file zipped backup archive of a project.

Most users do not have a deep enough feel for the differences and nuances here to take all the right steps before, during, and after a file loss event in order to recover their data without the use of an appropriate backup technology that is designed around then once or of point-in-time backups. Whether that’s Time Machine, Windows Backup, an external disk, an external SAN, a cloud backup service, or (ideally) some combination thereof is up to each user to figure out.

Advanced users who understand the technologies involved and are comfortable with the implications can always take shortcuts or use pieces creatively, but it isn’t just advanced users reading these forums and asking for help and following the advice here. We owe it to them to be very clear about how we use these solutions and about the hidden expectations necessary to be successful with those solutions, or the advice we give will inevitably cause damage.

1 Like

Dropbox does have versioning, and I’m not sure point-in-time copies is the universal definition of a backup. That said, I do have other backups that fit your definition, and of course it’s a good idea.

Does it version the entire session (all of the various folders and component files) at once, or does it version each file independently? (EDIT: like OneDrive and other sync services, it’s a per-file version system with no concept of file dependencies.)

I’ve taken a lot of backups, run a lot of restores, and helped many customers backup and restore a LOT of different data and applications over my quarter-century career work as a systems administrator and IT architect. Unless you can successfully restore your data, it isn’t a backup. Manually fiddling with file versioning on sync engines is great for singleton files, but once you have dependencies between multiple files, point-in-time copies are the difference between successful restores and wasted bits.

  • We tell users they can sync between Mac, Windows, and iOS with Dropbox, so long as they close on one device before opening on the other. That can’t work if all files in the project don’t sync to the cloud, given a reasonable amount of sync time – and in my experience, it does work.
  • More than once, when upgrading to a new computer, I have restored Scrivener projects along with virtually everything on my hard drive except system files. That also worked.
  • A user may not allow enough time for a complete sync in the press of day-to-day opening and closing of projects (perhaps), but restoring from backup is a different use case. It doesn’t happen as quickly.
  • Beyond that, along with also using ChronoSync and Time Machine, in a pinch I’d first go to a zip backup in Dropbox. That’s one file and your cautions don’t apply.

I would not advise any user, no matter how knowledgeable, to depend on Dropbox versioning to recover a specific point-in-time version of all or part of a Scrivener project. I’ve tried it, and the Scrivener project format is too complex and what Dropbox is doing is too opaque. Dropbox versioning might work as a backup of last resort when all other alternatives have failed, but my recommendation would be to structure your backups so that you don’t get to that point. Saving Scrivener’s automatic ZIP backups to a different cloud service from the live project is a good start on a robust backup strategy.

2 Likes