Snubbers are simply a resistor and small capacitor with values selected to suppress the oscillations (aka ringing) from switching or the EMI emitted when diodes start and stop conduction. There is a method to determine the values for optimum snubbing by trying different capacitor values while watching the ringing on the oscilloscope. A 50% reduction then lets the engineer calculate the optimal values. This is only a modest reduction in the ringing or EMI, and if you use your trusty AM radio near any wire connected to your GTI, you will see that there is a whole lot still left. The values used for this are small, and excess capacitance just causes power loss in that the power transistors must overcome the capacitance when switching on. Values of 0.001 to 0.1 uF and R values of 1-100 ohms would be ballpark ranges.
For simple on-off controls that aren't pulsing rapidly as in a PWM controller (say at 50,000 times per second) they are often not used at all.
When turning off an inductive load, there is also an inductive kick back which can cause some issues- if the low side (negative) is being switched, when the load is off, the voltage goes high. It can be more than the switching transistor(s) are rated for. Here a small capacitor can absorb some of that spike and keep the voltage within safe limits. Selecting the transistors with voltage rating of near double the actual use will often provide margin to be able to ignore "snubbing". The value for the cap is small- 0.01 to 1.0 uF. AC or DC are handled differently, and there are several techniques commonly used for DC relay coils, a notorious inductive spike generator when switched off due to the very high inductance. The techniques may include both a diode and snubber. If the switching is infrequent, and the voltage rating of the transistors is high enough, no snubbing is needed.
Slower switching, with limited slew rate also reduces this inductive spike problem and this is the technique I use for switching DC electric cooking elements along with picking transistors with a fair amount of head room- typically 250 volt rating for switching 120-140VDC. I use no snubbing and limit the power transistor on/off times via high value resistor to gate...this can only be done when using MOSFETs rated for semi-linear operation.
AC relays, switches and breakers on DC typically fail on opening. DC rated relay contacts must be way tougher, they must open faster and further. Bouncing of contacts is normal so each open or close is actually a series of events...and even closing causes arcing. DC also means that metal will always be transferred in one direction between contacts, which is very hard on the contact life. Adding capacitance will increase the current on closing of contacts but will reduce it on opening. Arcing on opening will be helped by capactance since the difference in voltage seen at the contacts at opening will be smaller, and the current across the open contacts will be reduced somewhat. So capacitance can improve things. But I would not normally use AC relays, switches and breakers in and around the home on high voltage DC. They WILL fail.
Even my 150VDC rated wall light switches fail over time. Lights with regular daily use of say 10 times a day will start to fail (sound of arcing on switching) in about 5 years. The old design rotary lamp switches also start to fail on DC in time, the ones most often used, again in about 3-5 years. I have metal electrical boxes and all wiring in metal conduit, so there is no risk of fire. The solid state switches (HV mosfet) in my cooking appliances have never failed...solid state switches are the way to go for DC.
DC mechanical switching is very different from AC; AC has spoiled us because the interruption of current flow every 8 or 10 milliseconds (for 60Hz or 50 Hz power) means arcs stop almost immediately, so all mechanical electrical contacts have a much easier and longer life. They are "value engineered" to just work well enough for AC. The exception is rotary lamp switches which are still the same as the original DC version. I also did find one double pole switch that has worked well for 120VDC at up to 3 amps ...but it is no longer in production, and I blew over $200 on testing others to find not one that would work.
"What I was wondering was if having a constant small load would help with the arcing?
My idea which I'm sure is way too simple to have legs, was to have something like a pair of 25W globes in series ( for voltage handling) across the switched terminals of the relay so effectively there was power to the heater element all the time.... Just very limited and inconsequential.
If the circuit was not completely dead, would that give the power a path to follow and reduce/ eliminate the arcing when the relay switched?"
The parallel small load would reduce switching current and voltage only very slightly. Not enough to allow use of AC relays and breakers. Some AC breakers are said to work for 12/24V DC, but since the breakers sold by Midnight Solar are DC rated and approved (150 and 300V versions), it seems foolish to tempt fate. Even with breakers, it is recommended that for an off grid system, a DC rated fuse of high rated current be used between battery and breakers since a sudden dead short could cause instantaneous currents so high (before the breakers can open) that the breakers contacts will weld closed.
DC is much better (4x) than AC for shock safety, but greater caution needs to be taken regarding mechanical switching contacts.
"If I put a lamp between the AC side of the inverter and the supply, would that allow the inverter to sync with the grid and stay connected?"
Alas, no. The inverter will immediately see an overvoltage when it tries to feed the circuit due to the resistance of the lamp. Same as if you used a long run of tiny wire between the GTI and power panel.