[R] installing R packages in a networked lab?

[Prof Brian Ripley]

Michael,

It is very likely that a network install will not work.  A security patch 
ca 2 years ago to Windows XP (etc) disallowed loading .chm files from a 
network, and you can stop loading .dll files too (and Vista does now by 
default AFAIR).

What we do is to install R locally (via the .msi installer that can be 
made: see the R-admin manual) and have other packages on a network drive 
*with* the chtml directory deleted*.  That way students get CHM help on 
the basic packages and text help on those loaded from the network.

However, you want to 'freeze' an image (or your sysadmins do).  I would 
build a custom installer including the packages you want and hand it to 
them to install locally.  The R-admin manual tells you how to do this: it 
may well be daunting but I remember you work with John Fox and he has done 
this (and will probably be willing to advise you).

Final comment: I suggest you do this with R-patched.  It looks like we 
will not have an R 2.7.2, but R-patched is now in a releasable state, 
precisely because people will be wanting to install something for the 
coming academic year (North Hemisphere).

On Fri, 8 Aug 2008, Michael Friendly wrote:

> Our support staff will be installing R 2.7.1 on a Windows Novell network 
> where R will reside on the server
> rather than on the local lab machines or at best will be in an image that is 
> copied to the local machines.
> I want to suggest a set of additional packages that should be installed as 
> part of the image beyond
> the base and recommended packages. 
> Once this is done, the image will be frozen for some time,
> so I want to make sure it gets done right. Moreover, the support staff who 
> will do this are not familiar
> with R, so I need to make it as easy as possible, telling them to do steps A, 
> B, ...
>
> I can think of two methods, but there isn't an easy way for me to test them 
> out, so I'd appreciate any
> advice.  In particular, I want to make sure that all compiled help files 
> (.chm) are also available.  Some
> of these, like rgl, will have .dll files in a ../libs directory. 
> 1:  Copy the packages I want from those already installed in my 
> C:\R\R-2.7.1\library to a folder
> on the server and tell them:
> A.  copy those folders to \\server\R\R-2.7.1\library
> B.  start R, and from the R Console, choose Packages -> Update packages...
>
> 2. I download all the Windows binary packages (.zip) I want, copy them to a 
> folder on the server, and
> tell them
> A. start R, and from the R Console, choose Packages -> Install packages from 
> local zip file(s)...
> B. select all, OK, have a coffee.
>
> what is most likely to be fool-proof?
>
> thanks,
> -Michael
>
>
> -- 
> Michael Friendly     Email: friendly AT yorku DOT ca Professor, Psychology 
> Dept.
> York University      Voice: 416 736-5115 x66249 Fax: 416 736-5814
> 4700 Keele Street    http://www.math.yorku.ca/SCS/friendly.html
> Toronto, ONT  M3J 1P3 CANADA
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595