That's 24:1 for total primary? And, 8 ohm load, so that'll be 4.6kohm a-a. In class AB, each tube sees 2.3kohm for the initial class A range, decreasing to 1.15k as the opposing tube goes fully into cutoff.
Seems pretty low, for two tubes and such a stupendously high supply voltage...
Yes, a 16 ohm load would probably be better suited, of course that will likely sacrifice some low-frequency response. Unless the transformer is in fact designed for 9.2k to 16 ohms, in which case there you go, your load is wrong and that's that.
6P3S is rated 20W or so, so a 15W idle is 28mA at 530V, or 44mA at 340V. In class AB, the peak current will be maybe 3-4 times higher, or 176mA tops (assuming enough screen voltage to draw that current at Vg1 = 0). 176mA into 1.15 to 2.3k is 200 to 400V. Maybe 300V mean, since we're going between the conditions in class AB.
530V seems to be an incorrect supply voltage.
8.4H and Q 19.1 is very believable for a choke of that type, that's good.
8.4H resonates with 4uF at 27Hz, so it's not a resonant type filter.
Likely, if the transformer and choke were original and as intended, it was supposed to be a choke-input filter (which gives 0.9 * Vrms or 340V DC).
8.4H at 50Hz mains, full wave rectified, has a critical load of 7.56kohms, or 45mA. That's one tube at idle, kind of marginal but should be okay. Load shouldn't be much less than the class A setting so this should be okay.
The supply voltage will swell very high (530V) until the tubes come up.
So yeah, remove the 4uF, and increase the other cap if ripple is a bit irritating.
Tim