auto download gone haywire?

cptwomey

My university library lost access to a major journal publisher who shut off access when my Zotero seemed to go crazy and submitted about 120 discrete requests for articles in about 3 minutes (most but not all were already in my Zotero library) to the publisher. The publisher assumed my computer had been hacked. It cut off my whole university's access to all of their journals, temporarily. My university library tracked it back to my computer.

My virus software is up to date, although I'll run more scans.

I'm curious: has this or anything like it happened to anyone else? Perhaps the new “Retrieve Metadata for PDF” feature may have spawned this somehow? Or perhaps ZotFile?

Many of the requests were duplicates. 16 different articles (13 of which are in Zotero, but 3 appear not to be).

Zotero was running in the background, but I was working on a slide deck, so wasn't interacting with it at all.

I'm perplexed and welcome any thoughts.

dstillman

Can you say what the publisher was?

No one else has reported anything like this, and if it was from Zotero itself it almost certainly didn't really happen when you weren't using it — Zotero doesn't make any non-zotero.org in the background. If these were items you saved recently, it's possible that something could've gone wrong during the saving of those items, but that's not something we've seen.

The automatic metadata retrieval, if that's what you mean by "auto download", only makes requests when you add PDFs, and only makes requests to Zotero servers, DOI providers, and ISBN providers, not publishers.

cptwomey

It was Taylor and Francis. If Zotero doesn't check in the background, that rules it out as the culprit. Would the browser connecter (to Safari) have generated repeated requests for data to the publisher's server?

I ran a few virus and malware checks and came up clean.

Probably something odd about my set up if no one else has had this problem. But I'd like to narrow down the source (at the very least, so I don't get the library mad at me again!) :-)

adamsmith

So the last time this happened it was because Zotero imported all items listed on a page rather than a single item (I forgot the details and no one got blocked from a publisher, but think importing the whole reference list instead of a single article). That could certainly have happened either because of a metadata error on their part or because Zotero did something wrong, but it's very hard for us find out retrospectively how. That said, you should have seen that while it happened, i.e. a pop-up with all the titles would have appeared when you saved from the article page.

Just to be clear, the connector wouldn't do this in the background, this would all be triggered by an import attempt.

cptwomey

That makes sense. I recall seeing that functionality (import whole ref list). That's a feature, not a bug. :-) But not what happened in this case.

OK, it must be something outside of Zotero. I'll look for clues if it recurs (hasn't in the past week).