Add --no-restart flag

[CriesofCarrots]

I want to test things like booting from a specific snapshot, so it would be handy to have the old /net ability to turn off automatic node restarts. startup_scripts.rs already supports a --no-restart flag.

I'm open to discussion on the best way to do this. It would actually be nice to have this configurable per node, but it serves my purposes for now to have a single hammer. Wdyt?

[gregcusack]

i think what you have here works for an all encompassing flag.
In terms of configurability per node, would you want something like ... --num-validators 10 --no-restart-count 4 ... or something, where 4 of the validators will not restart on failure? We could also add a label to the validators, so we would know which validators are being run with --no-restart?

Or would you like something that applies --no-restart to bootstrap, validator, and rpc groups of nodes individually?

[CriesofCarrots]

Or would you like something that applies --no-restart to bootstrap, validator, and rpc groups of nodes individually?

I have been pondering if, down the road, it makes sense to enable per-node configuration via a yaml file or something like that. Other configs could include amount to node stake and even software-version/build-type. The benefit of up-front configuration would be enabling setup that needs to happen at genesis for a heterogeneous cluster, like #66.

But given the capabilities already provided by the heterogeneous-cluster approach, I think an all-or-nothing --no-restart is actually fine for now.

[gregcusack]

yesss i think this is the right way to do it down the road. and to be honest, I probably should have taken the yaml approach for heterogeneous clusters from the start.

[gregcusack]

lgtm! thank you!