Disaster Narrowly Averted – A reminder to backup your files.

Last night, I was plugging away writing a new scene, when I got a very strange error from Scrivener.  I wish I’d taken a screencap of it. I certainly will if it happens again. The gist of the error was that Scrivener could not save the project. Those who use Scrivener will recognize that such a condition is exceedingly rare.

Scrivener is set up to autosave your project whenever it detects that you are idle for more than 2 seconds. It was during one of these autosaves that I received the error. I suspect that the way I have Scrivener set up may have been to blame. Furthermore, I don’t think I’m the only one with Scrivener set up this way.

For my current configuration (which I’m going to change), I have Scrivener saving my project file directly to a folder on Dropbox.  I wasn’t too concerned with file corruption or out-of-sync issues from simultaneous access across multiple systems as I only use one system to write. Dropbox sync provided me with a pain-free offsite backup solution and an automatic backup whenever I turned my other systems on. However, I suspect that Dropbox, like other cloud-sync services, must lock the files very briefly while it is performing its sync operation. I’m guessing Scrivener just happened to be trying to autosave just as Dropbox locked the file.

I’ve been using this configuration with Scrivener for about a year and I’d never run into this particular error before. Recreating it might be difficult.  In any case, when Scrivener gave me the warning message I realized that the file didn’t save correctly, I hit “Save As…” and kept on going. I received a number of other warnings at that point about Scrivener not being able to save and realized that something much more serious was going on. I closed out of Scrivener and reopened the project.

The scene was right where I’d left off, so I kept right on plugging along. I hit my word count and saved the file.  That’s when I noticed something was wrong.  Instead of the file being about thirty megs, it was about 30K! I reopened the project and checked.  The scene I had been working on was intact, but the entire rest of the novel was gone. All my inspirational photos, maps, research, deleted scenes, snapshots, all gone except for the new scene I’d been working on.

I panicked.

But I have backups. Lots of them. (Check out this post to get an idea of what I do). Most importantly, I have my Scrivener set to make an automatic backup when I open a project and again when I close it and I also have it set to back up every time I do a manual save. I also use Scrivener’s “Back Up To…” function to my Google Drive whenever I think about it.

These habits saved my bacon. The most recent backups didn’t help me because they had dumped the whole project as well. But the backup Scrivener generated when I had opened the project at the start of my writing session was good. All I had to do was cut and paste my new scene into the project, save again (and backup), and I was good to go.

I shudder to think what would have happened if I hadn’t been using Scrivener or if I hadn’t had my backup settings configured. Hence this post, since I figured some folks out there might not have things set up in a good manner to allow for easy data recovery in the event of something catastrophic.

Here are some recommendations:

• Make sure you have Scrivener backups turned on.  At a minimum, set Scrivener to backup when you close the project.
• Ideally, the backup location should be a different location than where you are saving your project file. Not just a different folder. But a different hard drive. Personally, I’m going to reverse my setup.  Currently, my project file saves to Dropbox while my backup files save to the local hard drive. I’m going to set it up so the backup files save to Dropbox and keep my project file on the local drive. Hopefully that should clear up the sync issue since the backup files will need to sync far less frequently.
• Compress your backup files into .zip.  It keeps things neat and will also help any cloud solution you have sync more efficiently.  Instead of trying to sync a folder structure, it will only be syncing one file.
• Use Scrivener’s “Back Up To…” function and set the location to yet another drive or cloud service. Myself, I’m using my Google drive account for this. If you’re a Mac user, I think you can set this up to save to this alternate location automatically. We on the PC have to do it manually, but it is still only one or two clicks.
• Once a week (or more frequently), compile your entire project into another format like .doc and save that to multiple places.

These steps won’t ensure that you can never lose your data, but it should make any single data loss event a lot less painful. For instance, for me to lose everything (and trust me, I’m not trying to tempt fate here), I would have to suffer a hard drive failure and two separate cloud services would have to go down.  And that’s not counting manual backups to flash drives, automated hard drive backups like Ghost or TimeCapsule, or services like Carbonite.

If anyone out there has tips on how to keep things more securely, more efficiently, or make it more robust, please comment.

UPDATE: Discovered this over at EndlessRealms.org.  I think the conflicted copy thing is what happened to my file. But instead of not being able to open it, Scrivener was unable to save.  In any case, it is more data, YMMV.

“Unfortunately, these conflicted copies can break Scrivener projects, which relies on files being in the right place. Scattered conflicted copies makes Scrivener think something is wrong and it refuses to load the project, claiming the project file can’t be found.

If you don’t have a recent full backup, here’s the way to correct this: Delete every conflicted copy. In some cases, if you have made different changes on different computers, you might have to decide which version of a file to keep in the Files/Docs subfolder, but other than that there should be no files named ‘conflicted copy’ remaining. You may find, if you have moved documents in the structure, that you have to re-move them, but that’s a tiny inconvenience, all your work is retained and nothing is lost. Once all the conflicted copies are deleted, you should find that Scrivener can load the project without problems.”

I have not yet tried this.  Also, it sounds like this might not work for folks on a MAC since you might not have access to the folder structure to drill down and get rid of the conflicted copies.