exporting NFS file systems with root permissions

[Brendan Eich]

In article <9006291911.AA25941 at avelon.lerc.nasa.gov>, fsfacca at AVELON.LERC.NASA.GOV (Tony Facca) writes:
> 
> I have a question regarding exporting file systems with the -root[id] option.
> On the Iris 4D's I have an /etc/exports file which looks something like this:
> 
> /usr/lerc -root=lerc08:lerc09,access=lerc02:lerc03:lerc04:lerc08:lerc09
> 
> This works the way I want it to.  That is, if I mount the /usr/lerc directory
> from lerc08 I can be root in the mounted directory.  
> 
> I am trying to duplicate this effect on the 3030's.  There is no "root" option
> but there is a "rootid" option which allows one to specify a userid to map 
> client "root" ids to.  Like, rootid=guest would change the permissions of the
> remote root to guest on the server.

The -rootid option was SGI's best guess and first implementation of what
Sun later called the -anon option (we implemented a root-uid-mapping option
before they did, based on incomplete hooks in their NFSSRC3.0 release).

The last release for the 3000 series, 3.6, was released a year before SGI
integrated exportfs and the -root option from Sun's DNFS1.0 source release.
The -root, -rw= (read-mostly), and -anon (synonymous with -rootid for SGI)
options are therefore available only on the 4D series.

> So, what exactly does the comment about "Lines beginning with white space.."
> mean?  

The comment lies.  To continue a long logical line across several physical
lines, end each line but the last with a backslash ('\').

Brendan