From: Jon Brinkmann
Date: Wed, 30 Nov 2005 15:13:22 -0700
Subject: Re: 2.5m obslog 20051129 (MJD 53704)
> >         We did make some headway on the mysterious grabInst failures.
> >         It turns out that when anyone except observer attempts to
> >         grab the spectrographs, it fails and in such a way that no one
> >         else can grab them either until that original session is quit.
> >         See PR 6782.
>
>     Excellent!!
>     Indeed, fuser, etc. can't investigate other users' processes (it
> wants to look at /proc/$PID/fd, etc to match the port number to the
> process). You probably want to stick to running spectro as observer
> only for this dark run. With luck a not-hopelessly-insecure solution
> can be put in.
>
>   - craig
This is a SUID issue for fuser.  Looking at sdsshost:
sdsshost % ls -l /sbin/fuser
-rwxr-sr-x    1 root     sys        18176 Feb 17  2000 /sbin/fuser
and sdsshost2:
sdsshost2 % ls -l /sbin/fuser
-rwxr-xr-x    1 root     root        18428 Oct 26  2003 /sbin/fuser
Should we set sdsshost2's fuser SUID bit?
Jon