LAM/MPI General User's Mailing List Archives

From: sean dettrick (sdettrick_at_[hidden])
Date: 2004-12-17 15:14:50

• Next message: Anthony J. Ciani: "Re: LAM: LAM 7.1 on Diskless cluster"
• Previous message: Reza Shahidi: "Re: LAM: LAM 7.1 on Diskless cluster"
• In reply to: Robin Humble: "Re: LAM: LAM7.1.1, OSX, usysv/sysv failing on some nodes, working on others"

> >through PBS (using tm boot) or using rsh boot. Meanwhile tcp
always works
> >perfectly on all nodes.
>
>typically people call an ipc cleanup script from PBS's epilogue script.
>

OK I wrote a script to hunt around and remove shared memory/semaphores. It
will clean up a crashed usysv job. But how can I be sure that I am not
killing shared memory in use by a *running* job? Any ideas?

Thanks,
Sean

• Next message: Anthony J. Ciani: "Re: LAM: LAM 7.1 on Diskless cluster"
• Previous message: Reza Shahidi: "Re: LAM: LAM 7.1 on Diskless cluster"
• In reply to: Robin Humble: "Re: LAM: LAM7.1.1, OSX, usysv/sysv failing on some nodes, working on others"