Print a failed command's error once on stderr (closes #433)

A command that failed after startup wrote the error twice on stderr:
as an ERROR log record and as the final ERROR: line. The failure
handlers now log the failure at info level, so the record appears only
under --verbose or --debug, next to the records that led up to it.

A failed blob upload is logged at info level for the same reason. A
restore that stops on a file it cannot restore no longer logs that
file's error, which the final line carries; under --skip-errors it is
still logged.

Model: opus-5-5

internal/snapshot/scanner.go

@@ -1875,10 +1875,12 @@ func (s *Scanner) uploadBlobIfNeeded(
 	err = s.storage.PutWithProgress(ctx, blobPath, blobWithReader.Reader,
 		finishedBlob.Compressed, progressCallback)
 	if err != nil {
-		// Only an interrupt cancels ctx, and the user is told about it,
-		// so an upload it stopped is not logged as a storage failure.
+		// The record is at info level: the returned error ends the backup
+		// and is printed as its ERROR: line. Only an interrupt cancels
+		// ctx, and the user is told about it, so an upload it stopped is
+		// not logged as a storage failure.
 		if ctx.Err() == nil {
-			log.Error("Failed to upload blob", "hash", finishedBlob.Hash, "error", err)
+			log.Info("Failed to upload blob", "hash", finishedBlob.Hash, "error", err)
 		}
 
 		return 0, false, fmt.Errorf("uploading blob to storage: %w", err)