How I see it:
Because there is a small amount of power leaking into the output capacitors, enough to fire the ignitor a couple times before diminishing, the SMPS is still running - at least partially.
But, because the power supply can be running without the laser itself running, I'd imagine there would be a way to switch off the tube voltage. Those MOSFETs, being IRFBG30's, rated at 1kV, 3A ea, may possibly be used to disconnect the tube from the power supply.
If they happen to be dead, it's possible maybe they're leaking the small amount of power into the output caps to fire the ignitor a few times. However, it's also possible the SMPS is disabled at a higher level, such as at the switchmode controller itself - I'm not sure.
However, I believe these are either 3, or most likely 4 layer boards, so it's possible an arcover may have occurred in middle layers - but not sure. If you can wipe all the soot off and it looks OK, then it probably was on the top layer, and the soot on the fiberglass kind of points to that also.
But I'd be testing all the semiconductors around that area - don't think it's all that likely for a passive to fail vs semiconductors on something like this.