Yes, 'pengguling' and 'gurami', they're all penultimate. Schwa is a good point, however. That's the one exception. The letter <e> can represent either /e/ or schwa, and the standard orthography doesn't distinguish. Etymological or paedagogic sources might mark them as é and ĕ respectively. It's never stressed . . . hmmm . . . or is it? I was going to say a word like besar "big" is phonetically [bə'sar], and is in fact phonologically /bsar/: schwa only occurs where unpronounceable consonant clusters would otherwise occur (including initial glottal stop followed by any other, as in enam "six" [ʔə'nam]). But looking down a wordlist, I notice words with penultimate schwa that I'm inclined to stress: terbang "fly (v.)", dengan "with". Maybe I've just remembered them wrong from Indonesian classes of many years ago. However, whether I'm wrong or not, still the only exceptions I can think of to penultimate stress are those words like besar, enam where the stress is final.
The first syllable of pengguling looks like a schwa-containing agent prefix of some verb guling.