Data at the charts are not precise, first of all, there are not many people I-CTS10228+ tested on the deeper subclades from Romania and Moldova; second, it is a big territory there might be regional differences: for example in anonymous researches in Western Wallachia (Olt, Dolj, Mehedinci) dominant subclade among CTS10228 is Dinaric South DYS448=19, which is also prevalent subclade of Western Bulgaria and East Serbia (so-called Shopluk) and also in Western Balkan. While in Calarasi region in Southeastern Wallachia majority of tested CTS10228 results are Dinaric North DYS448=20.
From public projects there are at least 7 Romanians with Romanian surnames tested as S17250>>PH908, one Dinaric North S17250, two people (including you) confirmed as S17250 without haplotypes, two people tested as S17250>>Z16971, plus 3 Dinaric North Y4460 and one Z17855.
The sample is not big, but very likely S17250 is a dominant subclade in Romania, it is also dominant subclade in general and make up approximately 60% of all CTS10228 and probably half of them are Dinaric South or PH908. Y4460 is the second common subclade with +40%.
While Z17855 probably takes about 5-6%, mostly appears in Balkan, especially in Central Bulgaria among speakers of the Central Balkan dialect
https://en.wikipedia.org/wiki/Central_Balkan_dialect, Greek Macedonia, Western FYR Macedonia, South Serbia and Montenegro. In my opinion, it will be a dominant subclade of Albanian CTS10228, too. Z17855 is to some extent present among Hungarians east from Danube and North Carpathians, so the guy who calculated the distribution of subclades among Romanians probably mistakenly includes tested Hungarian from Transylvania or just make a wrong prediction. Also among East Slavs (Russians, Ukrainians, Belorussians) and Poles Z17855 almost non-existent, so graf is very likely wrong.