[csw-maintainers] cswclassutils: location of template init scripts

[Philip Brown]

thank you for taking the time to understand the issue fully. more details below

On Wednesday, June 9, 2010, Maciej (Matchek) Blizinski
<maciej at opencsw.org> wrote:
>
>
> ..... IIUC, the NFS
> sharing is done in such a way that there is a master machine onto
> which CSW packages are installed and which serves the /opt/csw
> directory in a read-only fashion.  Other machines mount this directory
> via NFS and don't install the CSW packages at all.  It's enough to add
> /opt/csw/bin to $PATH in order to access the utilities.
>
> If you tried to install a package to a machine with read-only mounted
> /opt/csw, I'm guessing that pkgadd would fail (is that true?), because
> writing to the file wouldn't be possible.  If you don't run pkgadd on
> client machines, how are daemons started?  The "empty input
> cswinitsmf" idea above wouldn't solve that.

you would never "install a package" on one of these nfs client
machines. the empty input case is to solve two use cases.

1. when it gets automatically called in a zone, from a global pkgadd,
when the maintainer specified a file in the class under /opt/csw

2. to allow the script to be called directly, by the admins of those
nfs client machines, to "do the right thing" for demon setup on the
clients.