So basically there's the oxygen injector face which is sandwiched between the oxygen turbine housing on top and the fuel hot gas manifold below. This joint also holds the thrust chamber and the main injector face to the bottom of the hot gas manifold, below it. It's a series of bolts atop the threaded rods holding the stack together that ended up being problematic here. They don't put safety wire on them for some reason. In hotfires on the ground the preload on those bolts ends up releasing slightly. This was not a problem before because the rate that happened was slow enough to not matter whatsoever. It was a problem on Flight 8 because the rate that those bolts continued to unload was greatly accelerated by the then still-unsolved vehicle resonance problem. Eventually it was unloaded enough that the upper end of the seal of this whole joint assembly began to separate, allowing the extremely high pressure hot gaseous oxygen from the OTP above to mix with the hot gaseous methane from the FPB coming in from the side. This blew the engine apart (it was center engine E1, IIRC), and the direction that the joint began to unseal and eject hot gaseous propellant was in the direction of the neighboring RVAC, damaged that neighboring RVAC substantially enough to force it to shut down as well as damaging the lower pylons mounting the rvac to the ship skirt. the stress on the upper, central RVAC mount attaching it to the aft dome and thrust structure, from the center engine explosion and the subsequent unstable shutdown of the RVAC overstressed that mounting fixture and that's how that engine managed to completely liberate itself from the vehicle