Questions regarding migrate from OneDrive to Zotero Storage

kassiansun

I just purchased the Zotero Storage and planned to sync the files, and suddenly realize that it might have issue in my use-case, so I want to clarify these details before syncing in case I encounter any data corruption/loss later.

1. I only use Zotero on macOS & iOS, which means the data files in OneDrive is one-way, local -> cloud. I think this is fine (although not needed any more)
2. The attachment files are also saved in OneDrive, the key here is that not all files are saved locally, I don't have enough disk space. So my question is:
2.a. Zotero will trigger OneDrive download-on-demand to read the files and upload it. This should works fine.
2.b. But since I don't have enough disk space, that means once the disk is full, OneDrive will start to reclaim local files. Since the local files are still there, Zotero should be able to recognize them properly. Does this work?
2.c. [KEY PROBLEM] Zotero also support download-on-demand which is what I needed, how does Zotero handle this? For OneDrive, the files are visible in the file system, and file i/o will trigger the cloud sync and proceed after synced. For Zotero, does not-downloaded-yet files get deleted from file system? If that's the case, OneDrive will delete it from cloud. I think this is fine except that it triggers repeated add/remove in OneDrive in daily usage.

Assume the above theory, Zotero will do these steps once I turn on the attachment sync:
1. Download each file one-by-one and upload one-by-one (or at least in small batches)
2. Once the disk is full and OneDrive start to clean local files, Zotero should ignore these files as not-changed and will not try to read it again and download it again, so this should be a no-op at Zotero side.
3. Finally, I should have a almost-full disk with last synced files saved locally in OneDrive.

And if this is the case, how should I clean local files?
1. Delete them in Finder (and thus OneDrive)?
2. Free them in Zotero? I didn't see any such function.

And for the daily syncing, I want to know how it works:
1. Will zotero try to compare file hashes or file modification dates? If it compare file hashes, it will trigger repeat download from OneDrive
2. Can I use OneDrive as a in-place attachment file backup anymore? I think this depends on how Zotero handle the local space cleaning operation, but I don't need it to back up all files in OneDrive, even thought I'm working on and synced locally files should be a good start. (I'll buy a new Macbook with larger storage someday)

dstillman

You're saying you have your Zotero data directory in OneDrive? That's not a supported configuration, even if you were only using it on one computer, so you're kind of on your own here. I wouldn't expect file access from Zotero to necessarily trigger automatic downloading from OneDrive — we have to go out of our way to trigger that even for evicted iCloud Drive files — but if you're saying it does, that's good.

In any case, we can't speak to the specifics of how OneDrive will handle downloads and evictions during syncing, but one way or another, you'll want to sync and then move the data directory out of the cloud storage folder.

Once files have all been uploaded to Zotero, you can search for PDFs within your Zotero data directory and delete them directly. A future version of Zotero will provide options to automatically delete local copies of uploaded files in order to limit disk space usage when using download-on-demand.

kassiansun

@dstillman Actually I was thinking about deleting the whole storage folder, will it work, or this will trigger zotero to delete all files in cloud? So it’s more about knowing the mechanism that Zotero use to diff and sync files, it’s not as obvious as the SQLite database which has syncing state per record.

dstillman

You should do what I say, not delete the whole storage folder.

kassiansun

OK, it seems that zotero’s file sync logic is different from other file services, I’ll stay with OneDrive for a file system which supports local purging, good enough for me.

Long-term I’ll buy more storage and use both OneDrive and Zotero for backing up instead of purging, purging is a bit hard to align between different services, if I can’t use it now, it’s fine

dstillman

I mean, to be clear, you could just set up a smart folder in macOS that shows all PDFs within the Zotero data directory and delete all files in it every so often. But up to you.