Sorry for not responding sooner, I was out camping this weekend.
>At such -30C extreme cold, does one still follow the standard procedure of starting the stove with the canister upright and then inverting it after a minute or so once the preheat tube gets hot?
I had three inverted-canister stoves set up: the beta Caffin stove, a Coleman Xtreme, and an MSR Whisperlite Universal. All had set out overnight (low of -34C / -29F) with their canisters attached (so they wouldn't warm by handling), exposed to the air, and insulated from the snow with silicone pads. The Caffin stove had a new Snow Peak Giga Power 110g canister. I lit it with the canister upright (small flame) and _immediately_ inverted the canister, without waiting for the heat shunt to warm up. As expected, it flared up; I dropped my pot on the burner; and after about 20 seconds, the heat shunt warmed and it settled down to a nice flame. As Roger mentions in his post above, it was burning mostly propane at the start, which is why I immediately inverted it to save the propane for pressure. The wind had dropped so I didn't need a windscreen, and the heat from the burner warmed the canister a bit more, which kept the stove going without intervention.
The Coleman Xtreme had a Powermax canister that was about 1/3 full. It wouldn't start until I poured some (precious body-temp) water over the canister. It ran after that, but the flame didn't really get roaring until I heated more water to pour over the canister. If I had set the canister in a basin before pouring, I would have wasted a lot less water.
The MSR Whisperlite Universal had a 3/4-full Snow Peak canister, and it wouldn't start. I closed the valve, inverted the canister, and poured warm water into the concave base. I let it sit for a minute, so the heat from the water could work its way down the side of the canister to the liquid fuel, then lit it in fireball mode. It flared a while until the preheat tube warmed up, then settled down nicely. The water in the bottom of the canister soon froze solid, but it had done its job of producing sufficient gas pressure.
So two of the three stoves were non-starters (insufficient gas pressure in the partially-used canister to force either gas or liquid down the hose to the burner), and probably the only reason the Caffin stove started was that it had a brand-new canister. However, by simply applying a bit of warm water to their canisters, both non-starting stoves soon ran fine. I intentionally tried starting the stoves with chilled canisters, and inverted them cold, to see if/how they would work under these conditions. If I had either slept with the canisters or set them in water baths before attempting to start the stoves, the stoves would have had no problems starting and properly heating the preheat shunt/tubes, and thus I could have avoided the flare-ups from burning liquid-state fuel. I should also mention that I slept under a tarp, so the flare-ups weren't a problem; there's no way I would invert the canister without a hot preheat shunt/tube in the vestibule of a tent!
I also tested an upright-canister stove, but I'll post about that in a separate message.