I very much agree that this one is more worrying, so I'll talk about that first.
When a problem goes away and nobody knows enough about why that an answer would likely help anybody else, voting to close as off-topic with the "a problem that can't be reproduced" sub-reason is appropriate.
Often, when the OP solves their own problem, people vote to close as "a problem that can't be reproduced." That's wrong. When the solution might help others with the same problem, the question should be answered.
In the case of that answer, the OP's self-answer actually says specifically what to do to solve the problem. This might help others with the same or a similar problem. That question should not be closed as off-topic.
There's a secondary way "a problem that can't be reproduced" gets used, and that is to close old abandoned questions.
This is not bad or wrong, but I do think we should take steps to verify there's good reason to think they're really abandoned.
For this question, the OP's profile indicates they likely haven't visited the site in the last ten months. There's an answer posted but it's not clear whether or not it really answers the question that's been asked, in that it recommends something the OP said they already tried.
So that one looks good to me.
In this one, the OP was using Ubuntu 12.10, which was a supported release at the time the question was posted.
But there's been very little activity on the question since, and no answers. (Actually there is one deleted answer, but it's a me-too post and not a real answer, which is why it was deleted.)
In this case, the OP visited the site earlier this month, about twenty days ago. In such a case it's very appropriate to leave a comment asking the OP if they're still looking for a solution.
Is there such a comment there? Essentially yes--Seth commented, saying:
12.10 is now EOL (end of life). Please upgrade to a supported Ubuntu release (like 14.04) and report back. Thanks.
(It looks like the first four close voters voted to close without commenting to ask the OP if they were still looking for a solution. If that's true--I'm not 100% sure it is because I can't view exact close vote times, nor deleted comments--then that is far from ideal.)
Questions about end-of-life releases that are asked when the releases are still supported are generally on-topic, but what about when there are no answers? Basically nobody wants to go nuking answered questions for being about old releases. But do we consider questions off-topic as EoL just when they're new? Or anytime they're unanswered (in the sense of not having any answers that are accepted, have positive score, or otherwise appear valuable)?
I believe this is a gray area. Or, to put it more precisely: I believe this is an area of ambiguity that requires good judgment, and if we exercise good judgment, we may be better off having it ambiguous than having a set-in-stone policy one way or the other.
I think that in this case, in the end, proper good judgment has been applied. This is an old question with little activity and meaningful work on the problem within our site's scope would probably require upgrading or installing a supported Ubuntu release first. A comment was left, and the question can be reopened if the OP wants to keep working on the problem.
If there's nothing wrong with a question besides that it looks abandoned, and the OP is still visiting the site, we should comment and wait (a few days) for the OP to respond before voting to close. But in this case, there are additional reasons for us to want to close the question. So I think it was okay.