fix(mirror): ensure multiple runs (#9899)

1. After second execution, mirror starts to fail because of Primary key
constraints on the events table. Because mirror always took the the
first `system.mirror.succeeded` instead of the newest one
2. Mirror panicked during migration of fields tables

1. Adjusted the database query to order descending and limit 1
2. added missing assignment

- detailed logging if the copy from statement failed.

internal/v2/readmodel/last_successful_mirror.go

@@ -36,6 +36,7 @@ func (p *LastSuccessfulMirror) Filter() *eventstore.Filter {
 		),
 		eventstore.FilterPagination(
 			eventstore.Descending(),
+			eventstore.Limit(1),
 		),
 	)
 }