I think English something is sometimes pronounced thus:
My questions are what the labio-glottal stop is called, how it's transcribed, and how to transcribe its release.
After thinking about what you described, I was thinking and I think I realised you were probably referring to the colloquial shortening of the word that is indeed glottalised. I believe this sequence of sounds is best transcribed as some variant of [səʔm] with a pre-glottalised nasal.
It's a glottalized p, usually written [p']. But sometimes this term refers to ejective glottalization, with both closure and raising of the glottis, and of course in English we don't get raising of the glottis. So it's not an ejective.
As noted in some other answers, [sʌ̃ʔm̩] is a possible pronunciation of "something" for many of us. That's what you get when the labial closure for the final cluster is delayed until the [m]. But, though it's also possible, that's not the articulation you describe.
You are presumably hearing it right. "Something" is different from "rotten" in more than place of articulation: "rotten" = [rɑʔn̩]. "Something", in that pronunciation, is [sʌmʔm̩] (traditionally that vowel is transcribed with wedge unless it's in an unstressed syllable). Syllabicity is not a phonetic "fact" that can be decided by listening, so I included the syllabicity diacritic on phonological grounds. But transcriptions can be more phonologically-inspired, so if you have a reason to think that the glottal plus nasal is one segment, you can call it a glottalised nasal, yielding [sʌmm̰] and [rɑn̰]. A third option is simply [sʌmm̩] where the glottal closure you be be relegated to a phonetic detail rule. If the point is to be clear about that pronunciation, go with [sʌmʔm̩].
[Addendum]Putting an even finer point on it, you might contemplate [sʌ̃ʔm] vs. [rɑʔn], with or without syllabicity. The question is when the lips start to close, so a later closure would point to this transcription. You can't get even a near-minimal pair with labials, since in Standard English a.k.a. my dialect 'rotten-glottalization' only applies to /t/. But there are analogous examples with alveolar, such as "cotton" [kɑʔn̩] and "Canton" (Ohio, not China) [kænʔn̩], and in the latter I think [kæ̃ʔn̩] is credible.
Needless to say, claims about phonetic outputs require phonetic method to verify.
In light of Greg Lee's observation and focusing on the word "simultaneous", [p'] would not really be appropriate unless the stop also becomes oral. There are speakers of American English who are aggressive glottalizers and you get [slæp'm̩] for "slap 'em"; the question is whether the thing that you're talking about sounds like that.
In the [sʌmʔm̩] version, glottal closure may be initiated at the point of labial closure, but would not reach full closure. Phonetic simultaneity is actually gradient, and the supposedly "simultaneous" labial and velar closures of [kp] are not truly simultaneous, thus we would like to see just how simultaneous the closures are -- it's not really a matter for introspection, it requires measurement. Also, one must be careful about equating glottal approximation with full and sustained glottal closure. Very frequently, supposed glottal stops are lenited to glottalization in inter-sonorant context.
