When trying to schedule a volume, If longhorn instance manager is giving an error

failed to find metadata for snapshot <snapshot.img>

I know for sure the snapshot.img doesnt exist on any node anymore. is there a way to tell the volume or the replica of the volume to stop looking for these snapshots?

This is a broken state which Longhorn cannot recover. You might already have data loss. You can forcefully remove them by: 1. Find the location of

snapshot.img

file 2. Inside that directory, there are

*.meta

files which contains the information about the parent snapshot 3. Note them down to form a snapshot chain. Then manually edit them to remove the reference to the target

<snapshot.img>

Are you having backup for this volume? It might be simpler to restore this volume from backup