roachtest: always unstall on disk-stall failure

Previously, if the disk-stalled/* raochtests failed while the disk was stalled,
a defer would attempt to unstall the disk before completing the test. This
could fail if the context was cancelled. The stalled disk would then prevent
collection of artifacts.

This change updates the defer'd Unstall call to use a background context that
will never be cancelled.

Epic: none
Informs cockroachdb#102946
Release note: none

pkg/cmd/roachtest/tests/disk_stall.go

@@ -159,7 +159,11 @@ func runDiskStalledDetection(
 		m.ExpectDeath()
 	}
 	s.Stall(ctx, c.Node(1))
-	defer s.Unstall(ctx, c.Node(1))
+	// NB: We use a background context in the defer'ed unstall command,
+	// otherwise on test failure our c.Run calls will be ignored. Leaving the
+	// disk stalled will prevent artifact collection, making debugging
+	// difficult.
+	defer s.Unstall(context.Background(), c.Node(1))
 
 	// Wait twice the maximum sync duration and check if our SQL connection to
 	// node 1 is still alive. It should've been terminated.