I have finished my master thesis and now a couple of colleagues who were involved (one of which was the supervisor) are interested in "extracting a paper" from it. We have already written one paper on a topic covered in the thesis, but in a bit more detail. I am worried that this would be considered self-plagiarism since I have already done the thesis and am only really summarizing one part of the thesis.
It is plagiarism only if you do not properly cite it.
If you have not published the thesis itself, of course you can publish parts of it or summaries of it or even the whole thing unchanged. The requirement to avoid "self plagiarism" is to cite. For example, in the introduction write:
This paper is taken from my thesis [title] done in [date] at [school] under the direction of [advisor].
It's totally accepted to publish journal articles based on thesis chapters. For one thing, MS theses are not well abstracted or searchable. Even with Ph.D. theses, they are rarely looked at. Getting something into a journal article is doing the scientific community a favor. It's also good for you and your coworkers in terms of pub count.
Consider the opposite--should one not publish any articles during grad school (to save them for the thesis)?
Just cite the thesis. That's sufficient. Nobody considers this double publication.
If you treat the thesis (and the other paper) like you would any other previously published work, then you avoid self plagiarism. In other words, cite the thesis as the source of the ideas and quote from it as necessary. Since it is also likely that you hold copyright on it, though maybe not on the published paper, you can quote more extensively from it than you might if it weren't your own work.
Self plagiarism doesn't mean that you can't restate your own ideas. It means restating them without indication of the original. The problem with self plagiarism is that it makes it difficult for a reader who wants the complete history of an idea to follow it back to the origin. The reason it is a problem is that the original contains additional context that the reader wants. The original itself normally cites other work and contains other ideas that form the context of the idea under examination.
Ignoring the plagiarism issue and answering only the top-line question, yes, it is normal to extract publications from theses.
Adding to a few answers here, I agree that there's nothing wrong or unusual about this. My MS thesis lead to two conference papers and a journal paper.
It's usually a good idea to publish papers from thesis work for at least three reasons: first, papers in well known conferences or archival journals are much easier for other researchers to find, so there is a much greater likelihood that your work will be recognised and cited by others. Second, condensing your results from a thesis, which is usually roughly 100-200 pages, to a paper, which is usually 5-25 pages, depending on where you publish it, will make it easier for the community to digest your work. Few active researchers have the time to wade through a 100-200 page thesis, a great part of which will be a survey of basic concepts and results that they already know well, but almost everyone has an hour over lunch to read a 5 page article. Third, almost nobody reads thesis work because graduation timelines often make it incomplete or not fully peer-reviewed. While it's true that most theses must be defended in front of a committee, the background of the committee and rigor of the process vary considerably between institutions and departments. Submitting the research to a journal or competitive conference for review opens it up to criticism (very, very harsh criticism) from the wider community. There's a good chance that the remarks made by reviewers will give you a clearer idea of what's missing or could be better in your work. Although it's painful in the moment, in the long run it will make your research better and more likely to get picked up by others.