I had a closer look at the source, and suspect a patch like this might solve it, by running all scripts sequencially if the tty is strange.
Index: startpar.c =================================================================== --- startpar.c (revision 1898) +++ startpar.c (working copy) @@ -631,6 +631,7 @@ char *prev_level = getenv("PREVLEVEL"); char *run_level = getenv("RUNLEVEL"); char *splashopt = 0; + int notty = 0; (void)signal(SIGUSR1, sighandler_preload); (void)signal(SIGUSR2, sighandler_nopreload); @@ -813,8 +814,7 @@ if (tcgetattr(0, &tio)) { - perror("tcgetattr"); - exit(1); + notty = 1; } if (ioctl(0, TIOCGWINSZ, &wz) == 0) wzok = 1; @@ -904,6 +904,11 @@ } p->splashadd = calcsplash(num, argc, splashopt); p->num = num++; + if (notty) + { + interactive_task = p; /* no tty, treat as interactive */ + continue; + } if (interactive_task) continue; /* don't start this here */ run(p); Anyone able and willing to test if it solve the problem for OpenVZ? Happy hacking, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org