I know what the issue is. Sorry I didn't think of this yesterday.
The extinction model determines how much air (plus any combustion products present in the air) is needed to burn the fuel in a grid cell. It then looks to see if the total energy released by combustion is high enough to raise the temperature of that mixture above a limiting flame temperature. If it isn't, then combustion is not allowed. If the fuel is propane and we have pure air (no combustion products to dilute the oxygen), the flame temperature will be ~1800 C and higher than the limiting flame temperature of ~1300 C. Keep the fuel chemistry the same, but now tell FDS the heat of combustion is ~15000 kJ/kg rather than ~45000 kJ/kg. The same amount of air will be needed for combustion, but we are releasing 1/3 the heat. So our computed flame temperature will be about ~600 C and we will never burn the mixture. This is probably what is happening in your case, that the fuel chemistry you have defined doesn't comport with the heat of combustion you are using. If your overall fire scenario will be well ventilated that you don't expect extinction to ever occur, then setting SUPPRESSION=.FALSE. probably won't have a large impact. If you do expect extinction to occur in your scenario, you may want to revisit your REAC inputs and make sure they are correct.