Thanks for the speedy and explicit responses (I didnt expect such good
answers or nearly as fast). i apprecate the information greatly.
as for the shell thing, this is the C program i wrote to cheat a little
bit
#include<stdlib.h>
#include<stdio.h>
int main(void)
{
system("/usr/bin/lamboot -v /etc/lam-mpi/lamhosts");
return 0;
}
i compiled it and SUID'd it user nobody, this allowed me to run lamboot as
user nobody, i dont know if it is of any use to you (probably isnt) but
hey if it gives someone an idea....
alas i must hunt down a clustering system closer to what i need, so i
shall unsubscribe and stop flooding thy list with emails which serve no
purpose as you are doing good stuff here and i do not wish to detract by
being useless.
On Fri, 1 Aug 2003, Jeff Squyres wrote:
> On Fri, 1 Aug 2003, Jim Procter wrote:
>
> > I'm under the impression that LAM implicitly requires a shell session to
> > execute its demons and sub-processes, but...
>
> In general, yes -- LAM will require a shell in order to launch its
> executables.
>
> > > basically im looking to run something like apache (system service) but im
> > > not going to compromise my boxes security to do it.
>
> Please note that LAM is not intended to be used for system-administration
> types of tasks. If you're looking to do system-level tasks, LAM is
> probably not the right tool.
>
> > This is exactly what a GRID style authentication is for. So to rephrase
> > the question - does lamboot via globus authentication cause LAM
> > processes to execute jobs in a sandbox ?
>
> LAM uses the Globus fork scheduler. So, no.
>
> --
> {+} Jeff Squyres
> {+} jsquyres_at_[hidden]
> {+} http://www.lam-mpi.org/
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
|
