Did I lose my work? "Document could not be opened" Oh,horror

Dear friends: What happened??

My independently backed up Scrivener file shows the same error. Scrivener’s backup folder contains a working file from April 2nd. So, I’m missing one day’s work, April 3rd’s. It was a good day, I don’t want to lose the work. Can I recover it?

As suggested in the alert, I created a new project and imported the corrupted Scrivener project. In the new project, the imported documents are screwed up, namely the files I worked on yesterday:

I worked on file TITLE “A” and its contents are in a very corrupted and incomplete form in the file TITLE “G”.
The writing that should be in TITLE G, is in file TITLE A, in a normal manner.

So for some reason, the work I did on file TITLE A yesterday didn’t get saved properly and stored in an arbitrary other file.

If I right click the corrupted, newest .scriv project and select show package contents, I cannot open the most recent file; it says “the file isn’t in the correct format”. Also it shows April 2, as the date last modified, despite me having worked on it yesterday.

Scrivener didn’t save any of the work I did yesterday and corrupted my project.

WHY? What happened? How can this error be reversed?? And how can it be prevented?

Anybody? This is a devastating, atrocious Scrivener screwup!! I haven’t been able to recover anything. :imp:
We need to figure out the cause of this and how to prevent it.

If you haven’t already, contact the at the mac support email. There’s a support link on the main page that leads to it. I’ve observed in the past that they go through the queue of emails first, and then address any forum support questions that have no replies, then then look at any others if they have time left over.

mapp, can you open the corrupt file with a powerful text editor like TextWrangler? Maybe you can recover the text, if not the formatting.


Thanks for your replies.
I’ve just messaged them.
@ptram: Thanks, just tried this. TextWrangler was able to open the file (76.rtf) but it didn’t contain the newest version of the chapter. :confused:
It just showed what I could recover with Scrivener also which is an incomplete, corrupted version of the text.

It’s weird, from my cloud backups I downloaded a version of the project from earlier on April 3rd, while it was also corrupt and couldn’t be opened with Scrivener, I was able to open file 76.rtf normally, only it didn’t contain all of my work.

I’ve used Scrivener for a long time and I don’t think I’ve made a mistake. I really wonder what has happened.

Whenever anyone says “cloud” I cross my fingers. Do you use Dropbox or Sugarsync, by any chance? Using the cloud increases the risk of losing data, and I wouldn’t go anywhere near it, except to save zipped copies of files. It only takes one dodgy connection to create havoc. I know a lot of people seem to use it successfully, but I still wouldn’t chance it myself.

Anyway, I hope you manage to recover what you’ve lost.


I have saved my Scrivener project in an encrypted .sparsebundle in dropbox. But this is not what I was referring to.

My whole HDD is backed up in the cloud using crashplan.com. Fully encrypted. I have unlimited storage and changes are updated as often as I want. I believe I have it set on 15minutes updates. Old version are kept 1 month back, then 3 months, etc… all depending on my settings.

Now Martin, on what grounds do you “cross your finger” and assert using the cloud increases the risk of data loss?

I would advise you to read this thread:

literatureandlatte.com/forum … f=2&t=5295

It explains better than I can some of the problems.


By the way, I realise that I was guilty of writing rather too little in the previous post, and didn’t explain myself as well as I could have done. Blame it partly on the fact that I keep experiencing a glitch in my Internet connection that loses the post I have just written (ironically) so I’ve started to get a little “short” in my posts.

What I should have said is that I asked the question because a number of people in the past have experienced data loss while using Scrivener with SugarSync, which is known to cause problems, and also while using Dropbox in less than ideal ways. So it occurred to me that your problems could have been the result of similar practices. I wasn’t intending to mean that the cloud in general is risky, only that it is risky to use with Scrivener, unless one is very careful about one’s methods. Having said that, I wouldn’t use Scrivener and Dropbox even in the recommended manner. I would only use it to save zipped backups of Scrivener projects – and I seem to remember that Ioa said more or less the same, and he is about the best authority on Scrivener, apart from Keith. You are obviously doing something slightly different, but it wouldn’t surprise me to learn that using the cloud has contributed in some way to the data loss you have sadly experienced. But that is only supposition on my part. I can only say I’m sorry you’ve had a bad experience, and I hope you find a solution to the problem. And be careful how you use Scrivener with the cloud!

I got a nice email from Jeff. He doesn’t know what went wrong but suspects the .sparsebundle unpacking and encryption process is the culprit.

@Martin: Thanks for the link. I read it. A suggestion to the mods would be to include this warning in the Scrivener tutorial.
Also thanks for the rest of the message. I’ve stopped using dropbox now, although I’m not quite convinced it was the problem.

What I’m doing now and anybody’s comment is invited:

-Scrivener project in a folder on my desktop. No more sparsebundles and no more dropbox
-Scrivener settings:

*Autosave after 2 seconds of inactivity
*Backup: automatic backups, check. back up on project open, check. back up on project close, check. back up with each manual save, check. compress back up as zip files, check

-crashplan backups: since this is 1-way I’m not afraid of corruption. I’m keeping the same settings as detailed above

-daily backups on external HDD

EDIT: A possibility would also be to include the active Scrivener project from the crashplan backups and only have crashplan backup the automatic Scrivener backups. Thoughts?


After reading Martin’s post including AmberV’s linked one, I have decided to set up my Scrivener as follows:

My active Scrivener project is saved in a folder “BOOK” in my HOME Folder on the local HDD. For quick access, I linked it in the Finder under favorites: snapplr.com/kac2

Scrivener’s auto-saving is set to 2 seconds after inactivity

Scrivener’s auto backup option is set to back up on project open, close, and after manual saving.
Backups are saved as zip with the date in the name. I’m keeping the ten most recent backups.


I’m using crashplan.com for my cloud backups. I’ve researched all backup services and crashplan is the best.
I’ve excluded the home folder “BOOK” from the automatic cloud backups. This is to prevent any possible corruption caused by the automatic syncing.

I included Scrivener’s backup folder under Library/Application Support/Scrivener into the cloud backups.

So the active project is not getting backed up in the cloud but the automatic backups are.
I have the habit of not closing my Scrivener project after finishing working and just close my laptop.
I must STOP doing this because a backup file won’t get created if I don’t close the project UNLESS I manually save it!!!
This is important, I will discipline myself to always close the project when I finished working for the day and to manually save after a few hours of good work.


Thoughts? Feedback?

Your comments are welcomed.

EDIT: Changed it to keeping 25 backups.
I’m questioning the signification of having a backup created upon opening the project. Thoughts?

Overall, that sounds like a much safer way of doing things, but I would never leave a program open if I was not using it. If it’s not right in front of my eyes, it is wasting electricity, memory, and cpu, in my view. You can also run out of battery power, and the memory will be lost, which may result in data loss.

I’ve been using Macs since about 1993, so I guess I have some old-fashioned approaches to things, but I also believe in safety. I have a “shut-down routine” for the end of the day – I manually close projects, and manually quit all applications. I then shut down the computer completely. This means that when I start work again in the morning, the whole thing is “virgin”, the memory has been cleared out, all files are opened anew from a previously saved state, and so forth. I have also experimented with a Keyboard Maestro macro which does the same “shut-down routine”, but I find doing it manually is also a way of “shutting down” myself.

I also run Disk Utility periodically, and Onyx every once in a while, to carry out system maintenance and clean out all the caches.

Best of luck with it,

Not staff, but I personally turned backup on open off - the file it created would be identical to the one created on close, and would eat into your number of rotating back ups. I’d personally suggest choosing between back up on close/back up on open to maximise the rotating saves, but I may be missing some reason why someone would WANT two identical copies of all their back ups.