What to do if a multi-party computation fails?

[sbellem]

In the current system, each node listens to on-chain events, and for some events it launches a multi-party computation. To perform a multi-party computation, each node invokes an MP-SPDZ executable (malicious-shamir-party.x) with some arguments. This step is "blocking" in the sense that the program flow stops until malicious-shamir-party.x returns.

For instance (complete code link):

// ...

cmd = exec.Command(
	"./malicious-shamir-party.x",
	"-N", players,
	"-T", threshold,
	"-p", serverID,
	"-pn", mpcPort,
	"-P", blsPrime,
	"--hostname", leaderHostname,
	"hbswap_trade"
)
err := cmd.Run()
if err != nil {
    log.Fatalf("cmd.Run() failed with %s\n", err)
}

// ...

The call to log.Fatalf() will call os.exit(1) which will cause the program to terminate immediately:

Exit causes the current program to exit with the given status code. Conventionally, code zero indicates success, non-zero an error. The program terminates immediately; deferred functions are not run.

Could we improve this? Could the server let others know that: "Hey, had a problem with this computation ... we need to re-try ..."? Would it make sense to create an event on-chain, which could then used as a coordinating mechanism for the MPC nodes to react in a consistent/safe way with respect to this failed computation?