This is an interesting question, and I think it can productively be split into a couple of distinct parts, separating the publisher side from the author side:
Was/is there any pushback from journals about posting pre-publication versions of papers on the web?
Why do research article authors in astronomy post papers on arxiv at a relatively high rate?
Part 1 is easier to answer, since there are fewer individual actors to consider. As far as I can tell, the answer to this is a flat "no". As @ProfRob notes in a comment, there was a long tradition of dissemination of preprints on paper before the web existed, so it likely would have been difficult for publishers to forbid that practice once electronic dissemination became a possibility. Many journal publishers are directly connected with professional societies, and even for-profit journals need to show support for wide dissemination of ideas. I don't know of any journal that publishes astronomy research that doesn't allow posting of preprints.
Contrary to what one often hears, the higher-profile journals like Science and Nature are on board with this idea as well; both of them are very explicit in stating policies that allow both pre-publication posting on sites like arxiv, and free discussion of results at meetings. (current Nature policy, current Science policy). What they don't allow is authors discussing results with the press before publication. Nature in particular seems to emphasize the importance of preprints; they published their first editorial on it in 1997, and they seem to publish articles every few years that say, in effect, "no really, it's ok."
Part 2 is harder to answer, as is any question about why a practice evolves in some particular setting that involves individual decisions by many different people to adhere to some unstated norm.
First I thought I'd check to see whether it's really true that astronomy is different from other physical sciences in its use of preprints. The article linked in the answer from @Thomas suggests that astro is on the high side (at least as of 2011) in use of arxiv. I did my own spot check with a few ADS queries. To keep it simple, I just picked a few journals. In the years 2015-2019, this ADS query shows that there were a combined 24,634 refereed articles published in The Astrophysical Journal and Astronomy & Astrophysics (to pick a major American journal and a major European journal). Of these, 21,568 have preprints available on arxiv. (You get this by adding the qualification "property:eprint_openaccess" to the query.) So that's almost 88% of papers, which seems high indeed. To compare to another sub-field, I chose the journal Physical Review B (which publishes papers on condensed matter physics, since I know there's a cond-mat section of arxiv), yielding 25,952 articles, of which 16,936 have arxiv preprints. So that's 65%, still pretty high, but a fair bit lower than astronomy.
So, why might this be? My guess is that part of the answer might lie in the creation of ADS, and in particular the ADS project to scan all of the previous astronomical research literature and make it freely available online. That made it possible to sit at your desk and do all of your bibliographic research, rather than going to the library. Once people got used to doing that, it was natural to want to do it for the newer literature, too. Of course those electronically published articles were likely available to many researchers directly from the publisher, through their home institutions, but that early digital access was pretty heterogeneous by journal and sometimes not so smooth. That might have contributed to people wanting to put papers on arxiv to make them more available.
It's certainly more complicated than this, but I do think the existence of ADS (including its indexing of arxiv) may have made a difference in earlier / more complete adoption of electronic dissemination in astronomy compared to some other fields.
A related thing that I think is interesting is that the use of arxiv in astronomy has now become so ubiquitous that some researchers consider posting on arxiv to be important not just for early visibility, but for visibility at all. By that I mean that the daily arxiv postings have become, for many, the way that they learn about new literature, supplanting the role of journal table-of-contents services. I don't know how widespread the practice is (and I'm sure there are some generational differences), but I have certainly heard people say that arxiv is the only resource they use for regularly keeping up on new papers that are published. In other words, there's a sense in which simply publishing your paper in a refereed journal is not enough to make it fully visible to the community, at least not in the short term (though people will likely find it in literature searches later).
To check this idea, I went back to the ADS astronomy searches referenced above and looked at citation numbers. The 21,568 papers with arxiv preprints have 532,493 citations total, while the 24,634 papers overall have 561,738 citations. Subtracting, that means that the 3,066 papers without arxiv preprints have 29,245 citations. Looking at average citations per paper, that means that the arxiv preprint papers have 24.6 citations per paper on average, while the non-arxiv-preprint papers have 9.5 citations per paper on average. That is such a huge difference that it actually makes me wonder if there is some confounding factor I'm not considering - I encourage others to check the numbers and/or to try other searches. But if that's true, it's a huge effect, and bears out the idea that arxiv plays an important role in the community in research visibility overall, not just in early access.
Finally, I'll note that the importance of arxiv is now so complete that you even see an open letter from the AAS journal editors to arxiv, asking them to include the Research Notes of the AAS (a new publication of short, unrefereed papers) in the arxiv. There has been an ongoing discussion of this, since arxiv moderators started (mostly) rejecting arxiv postings of RNAAS papers.
