node_A: issued
node_B: pending
Fabric. Same certificate. Same mission hash. Different room, apparently.
Node_A had already carried it forward. Mission ledger updated. Certificate visible. The kind of screen that makes somebody say “okay, we’re good” before they should.
Node_B still wouldn’t move.
No missing proof. No broken registry entry. Just pending state hanging there while the faster node was already speaking in settled language.
People call that a sync issue like it stays small if you name it gently.
It doesn’t.
The next dependency edge got released off node_A’s view before node_B agreed the run was complete. That's when Fabric stops feeling like a ledger problem and starts feeling like scheduling built on a disagreement nobody can see cleanly.
Support asked for the mission hash.
Ops checked the ledger twice.
Someone said, “node_A has it,” like that should carry the rest of the room.
It didn’t.
By the time node_B caught up, the bad part had already happened.
The next step was already standing on the faster version of truth.