Bug #56125 MTR2 start-and-exit removes server tmpdir, server becomes not operational

This happens when creating new tmpdir due to too long socket path
Don't delete it if --start-and-exit, but warn user to do it.

mysql-test/mysql-test-run.pl

@@ -110,13 +110,25 @@ my $path_vardir_trace;          # unix formatted opt_vardir for trace files
 my $opt_tmpdir;                 # Path to use for tmp/ dir
 my $opt_tmpdir_pid;
 
+my $opt_start;
+my $opt_start_dirty;
+my $opt_start_exit;
+my $start_only;
+
 END {
   if ( defined $opt_tmpdir_pid and $opt_tmpdir_pid == $$ )
+  {
+    if (!$opt_start_exit)
     {
       # Remove the tempdir this process has created
-    mtr_verbose("Removing tmpdir '$opt_tmpdir");
+      mtr_verbose("Removing tmpdir $opt_tmpdir");
       rmtree($opt_tmpdir);
     }
+    else
+    {
+      mtr_warning("tmpdir $opt_tmpdir should be removed after the server has finished");
+    }
+  }
 }
 
 sub env_or_val($$) { defined $ENV{$_[0]} ? $ENV{$_[0]} : $_[1] }
@@ -215,10 +227,6 @@ my $opt_start_timeout   = $ENV{MTR_START_TIMEOUT}    || 180; # seconds
 sub suite_timeout { return $opt_suite_timeout * 60; };
 sub check_timeout { return $opt_testcase_timeout * 6; };
 
-my $opt_start;
-my $opt_start_dirty;
-my $opt_start_exit;
-my $start_only;
 my $opt_wait_all;
 my $opt_user_args;
 my $opt_repeat= 1;