How and where does a corda node (Y1) leave a message for a backup node (Y2) so that Y2 can resume the flow that got interrupted because Y1 went down?

Question

When in a hot-cold corda deployment where say the nodes are: c1 (PartyA), c11 (PartyA backup), c2 (PartyB) Now if a flow has been initiated by c2 and c1 is responding but it goes down while doing that, then where, if at all anywhere, is the message stored for c11 to read from and know what c1 couldn't complete and hence should be attempted again.

More context about the above setup -

1. There was no shared database setup (this can be the only other thing that c1 puts the info at for c11 to continue but then what if before committing even that it goes down)
2. I used an infinite while loop in the responder's call() method to simulate a stuck scenario

I used to think it was the artemis folder but after mounting a common network file storage in the artemis folder of c1 & c11, when c1 was put down c11 still did not attempt to initiate the responder again. Please do let me know if that while loop is an incorrect way to do things, I'll then change that first before continuing. It really maybe that a checkpoint is required to be saved for c11 to pick up and resume but then the question would be that what if c1 goes down before saving the first checkpoint.

Answer 1

等待大神解答