For-profit publishers (my experience is mostly with Springer and Elsevier) typically have some conditions on what you can an cannot do to self-archive your manuscript. For example, Springer allows you to self-archive on your own website immediately but has sometimes an embargo period for arXiv and similar public repositories.
As far as I understand, you can get around the embargo period by uploading the pre-review text (And I suppose not using the publisher's template is best).
Placing your articles on public repositories is a common practice, so I am wondering: what happens when inevitably someone steps out of line and puts on arXiv a version too close to the published manuscript, or does it too early.
Are there documented cases of publishers complaining? And do the publishers simply send a removal request to you/arXiv, or do things get heated?