I'm running on CentOS 5.7 (in case that matters) I have a program (not my own) that runs fine from Lonsole, but when I run it from xterm, gnome-terminal or rxvt, it just hangs.

I'm doing this with a completely fresh account, no .bashrc, no alterations to defaults. Brand new vanilla install.

On both terminals, $- is himBH so, both the same type of shell, I would think.

If I ssh into the machine, even from itself (ssh, it works in all terminals.

Any idea what could be wrong?

EDIT: The program is qconf, part of SGE (GridEngine), in case that matters. It goes to a server and returns a list of machines that are allowed to make compute requests.

I ran with traces, the diff looks like:

diff broken working
< execve("/opt/SGE/bin/lx24-amd64/qconf", ["qconf", "-ss"], [/* 49 vars */]) = 0
< brk(0)                                  = 0x1a48c000
< mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ba0bec8a000
< mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2ba0bec8b000
> execve("/opt/SGE/bin/lx24-amd64/qconf", ["qconf", "-ss"], [/* 48 vars */]) = 0
> brk(0)                                  = 0xa7cb000
> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =     0x2b3b4716e000
> mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2b3b4716f000

< futex(0x3a408030ec, FUTEX_WAKE_PRIVATE, 2147483647) = 0
< futex(0x6ba3c0, FUTEX_WAIT_PRIVATE, 2, NULL
\ No newline at end of file
> open("/proc/cpuinfo", O_RDONLY)         = 3
> read(3, "p", 1)                         = 1
> read(3, "r", 1)                         = 1

[working goes on from here]

Broken stops in mid futex output, which is odd.

  • What does the program do? Can we see it?
    – terdon
    Commented Jun 21, 2013 at 15:40
  • Do you run the program directly inside a terminal, or from a shell? What is the program doing? At least post a trace: xterm -e strace -o your_program.xterm.strace ./your_program. Compare with the trace in Konsole. Commented Jun 21, 2013 at 15:44
  • What version of GridEngine (SGE) is this?
    – slm
    Commented Jun 24, 2013 at 16:53
  • At the bottom of the man page it says 6.2u5 Commented Jun 24, 2013 at 17:56
  • One possibly interesting thing is that the final call to futex appears to have an extremely small 3rd arg, and an unnecessary 4th arg... Commented Jun 24, 2013 at 18:02

1 Answer 1


I have the same configuration i.e.: Centos 5.7 and SGE 6.2u5. No problems. I can work directly on server console, or x-term. I can ssh both in text- and x-modi. Everything works.

The only thing that comes to my mind is initialisation of SGE-environment. Are you sure it's done properly? I had some users that "somehow" managed to make it wrong. For that reason I added SGE-initialization to /etc/profile. Smth like this:

   if [ -f /Services/NFS/SGE/default/common/settings.sh ];then
     . /Services/NFS/SGE/default/common/settings.sh

Did you compiled SGE your-self or did you use Sun's binaries for 64-bit linux (ge-6.2u5-bin-lx24-amd64.tar.gz ge-6.2u5-common.tar.gz)?

  • I'm not admin on the system, so I didn't compile anything. I'm accessing the system via a virtual computer system called No-Window, or NX. It's possible that NX isn't set up correctly, and so it breaks things... But how would that affect SGE? Commented Jun 26, 2013 at 20:01
  • There could be a lot of reasons. Missing/incompatible libraries, wrong export of system variables for the start. But you can check it easily enough. NX is a remote desktop. SO I assume you work with this on windows or mac machine. Try to use x-server instead and see if this works. If you work on windows try the demo version of Xwin32 from Starnet. It works for 30min then it closes the session. Of course in text modus you can use putty from Simon Tatham.
    – Kris_R
    Commented Jun 27, 2013 at 9:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .