Zotero desktopwebdav sync uncertain failed upload; download, not upload working on app.

edited 3 hours ago
Hi.

Currently away from main desktop but found a bug that I thought was caused at my webdav server, but actually seems to be interacting between the zotero desktop (9.0.6) and app version 1.0.0-247.

I am currently using nextcloud webdav as the backend for my zotero storage, which was working really well up until recently.

I went to upload a colleagues thesis for later reading on the desktop version of zotero but found an issue. It's 10Mb in size.

1. The upload was causing a timeout error, and my crowdsec instance (that parses logs and traffic request) was getting bogged down and locking down the server. Unfortunately neither crowdsec, traefik, nextcloud or zotero was able to provide any meaningful insights aside from zotero getting a 404 when it runs a get request for the associated .prop file, and eventually causes crowdsec to crash so that everything gets a generic 403.

2. I deleted the offending file and changed the timeout settings on my crowdsec instance to be 60 seconds, it still causes everything to crash, but instead of crowdsec timing out and getting a 403 error when accessing the appropriate .prop file, instead zotero upload itself times out.

3. Today when working on my android the file was successfully downloaded from the nextcloud server, with no issues and synced just fine.

4. When I try to download the file on zotero desktop however it complains that the file doesn't exist locally, when I move to that directory I can see it downloads the right folder from the webdav, but its contents are truncated.

5. The folder is available on the webdav server, and contains the complete pdf, but there is no .prop file associated.

6. When I upload the file using the android app there are no errors on upload, however similarly when I try to download it, the zotero desktop complains that there is no local file, and the zotero web application cannot download or open the file either. It is not found on the webdav server also.

I'm kind of stumped tbh. I figured initially that maybe the parser for crowdsec was getting bogged down when using webdav, but that's not it.

Does anyone know a good way I can capture the headers and requests that zotero desktop and zotero app uses to communicate with the webdav server? This may help nail down why the pdf is uploading when done on desktop with no .prop file, and why it can be downloaded by the app, but not desktop, and why the app indicates successful upload or sync, but doesn't actually do this.

Thanks for any input :)
Sign In or Register to comment.