Conversation
| }) | ||
| }); | ||
|
|
||
| // TODO: https://github.com/Near-One/mpc/issues/441 |
There was a problem hiding this comment.
I didn't have time to look at everything yet, but if 441 is open, then what is the current mechanism to prevent including new participants in computations that require valid keyshares?
There was a problem hiding this comment.
It is fine. #441 is just asking to add an explicit step that checks membership of running state before spawning a task for the running job. Currently the spawned future just returns immediately with https://github.com/Near-One/mpc/blob/d186fcf87de34fc8157d2efa56b3a9bfb1d346a0/node/src/coordinator.rs#L87 so no work is actually being done.
There was a problem hiding this comment.
My question was lacking precision, let me rephrase:
What prevents the leader of a computation (e.g. triple, signature or pre-signature generation) from asking a participant that was not in the previous epoch (and thus does not have a valid keyshare) to participate in the computation?
There was a problem hiding this comment.
That's a good point. I updated the signature of select_random_active_participants_including_me to also take a set of participants to consider for the sampling 59aa238#diff-69623eb195b5eca6a57555b34713b6aca6491efab4f9920106f952acf8db4cda
I've also updated the participants to be considered by the running jobs to only be intersection of nodes that are in both thre running and resharing state so we don't consider nodes we are connected to in resharing that don't have active key shares.
There was a problem hiding this comment.
please delete and also delete the comment
kevindeforth
left a comment
There was a problem hiding this comment.
Thanks!
Could you please also update the pytests before merging?
| }) | ||
| }); | ||
|
|
||
| // TODO: https://github.com/Near-One/mpc/issues/441 |
Yes. Done now in a585241 |
| Tests that signature requests are still processed while performing key resharing. | ||
|
|
||
| Test scenario: | ||
| 1. Start with 2 nodes and 1 domain. |
There was a problem hiding this comment.
by default init_cluster uses two domains, one for ecdsa and one for eddsa. Both should be tested here.
| }) | ||
| }); | ||
|
|
||
| // TODO: https://github.com/Near-One/mpc/issues/441 |
There was a problem hiding this comment.
please delete and also delete the comment
No description provided.