Is there any potential impact of updating to Cupertino-17.9.7?
Well yes, hopefully it will be "better" than an earlier version. Unfortunately, that's not 100% guaranteed.
I assume your question really is, given a choice between a recommended (gold star tagged) version, and new subsequent, and sequential, release, which should be preferred?
However, if "the policy" is to use the "latest" Cupertino release (". . . I need to upgrade to the latest version 17.9.x but Cupertino . . ."), that would be .7, correct? Unless you're questioning that policy, or setting such a policy, your question is moot, isn't it?
Policy requirements aside, this really is a difficult question to answer. I assume you're trying to obtain the most stable version of Cupertino, and, again, there's no guaranteed way (as of yet) to determine that, answer.
Heck, the whole point of a recommended version is, it's the vendor's "best of the litter" opinion, but, interestingly, in this case, you note the recommended version is .6a, and I recall (?), lettered suffix releases are made when there's a bit of an "oops" with the published release. I.e. some issue(s) found, significant enough, that a correction for .# could not wait until .#+1 (or even another suffix letter release is needed, e.g. #a, #b, etc.).
However, if this is a choice you can make, and perhaps, one that will help refine a policy, some thoughts . . .
I would recommend reading the release notes for the .7 release, and see if there are any fixes made to it, that would worry you if known to still be in .6a. Also noting, how many changes were noted made in the .7 release. If the former, you would probably want to go with the newer .7 release. If few changes were made to .7, that possibly reduces the chances of introducing new bugs, you might want to use it, as it's very likely to become the next recommended release.
If neither is the case, next I would look at how long the .7 release has been publicly available. If it just came out, it's less likely it has been widely deployed by customers, so its quality is yet to be customer confirmed. So, you might want to just install the recommended version, and treat it, .7, however you normally would follow-on updates.
Also, in general, and if possible, any new software installation, should likely have extra scrutiny for new issues, for sometime; be a phased rollout, both in numbers of platforms updated, overtime, and potential (i.e. particularly sensitive) "location" impact, if defective; and plan for a quick roll back.
BTW, @Jens Albrecht noted, vendors, such as Cisco, often do monitor customer feedback, in making recommendations. But, what we aren't often told, is how many customers are using the product. Further, vendors, themselves, often don't know, the details of the customer usage of their product. So, customer feedback, although valuable, itself, also doesn't guarantee software is stable.
Another issue, is the "catch-22", i.e. customers may be reluctant to run software that's not "tagged" as stable, but if customers don't run non tagged software, we don't get widespread customer usage feedback that helps us determine quality.
Personally, I can attest to, when I would run into software bugs, it often was in using "exotic" features, i.e. features not widely used. Or, read the known bugs release notes. Many of them probably are of the class, "what's that for".
In theory, newer maintenance release should always be better, but in practice, that's not always the case, but, odds usually favor the former over the latter.
