[BioC] Warning message while trying to install Bioconductor

[Steve Lianoglou]

Hi,

For the record: I was not offering advice on the best way to install
and maintain R on a single user or multi user setting.

I only suggested to install as admin because my "reading between the
lines" lead me to guess that the OP was most likely using a personal
machine (ie. not working on/administering a multi-user machine) and
also had limited sys admin experience so painstakingly outlining an
admin strategy was likely not what he was after..

There are much better/correct ways to install R in these situations,
as you all are rightly pointing out, however my impression was that
the OP just needed to give the proverbial TV a swift knock on its side
to shake out the static in the picture and I was just pointing to a
spot that could be whacked so he could get on with viewing the
intended show (apologies to those of you who are lost by now who have
only ever known TVs to be of the plasma/LCD variety)

If I were giving any type of "official" advice (from my point of
view), it'd probably be something similar to the 3 tiered approach
Kasper outlines, but I'd encourage people to make use of a versioned
R_LIBS_USER directory to maintain their own updated version of
packages (even replacing (by precedence) the same package that might
be installed by your "group 2").

-steve

On Mon, Jun 2, 2014 at 6:33 PM, Kasper Daniel Hansen
<kasperdanielhansen at gmail.com> wrote:
> The warning is appropriate.  Wolfgang and I are disagreeing with Steve's
> suggestion that the OP should just update/install as administrator.
>
> Best,
> Kasper
>
>
> On Mon, Jun 2, 2014 at 3:15 PM, Martin Morgan <mtmorgan at fhcrc.org> wrote:
>
>> On 06/02/2014 07:28 AM, Kasper Daniel Hansen wrote:
>>
>>> And as a followup: what I do in a multi-user environment (and after some
>>> years of experience I think this is the best approach) is to immediately
>>> make a site-library directory where I install all packages.  Basically we
>>> have groups like
>>>    1) me
>>>    2) people allowed to install R packages system-wide
>>>    3) most users
>>> I set the permissions of site-library to be 2), but the standard library
>>> is
>>> always 1) because it is installed as part of R.  This mean any user in 2)
>>> who (say) tried to update all installed R packages system-wide will be
>>> told
>>> that the packages included with R cannot be updated; exactly the situation
>>> the OP described.
>>>
>>
>> I went looking in the R-admin manual for best practices along these lines,
>> and was a little surprised not to see anything. Did I miss something?
>>
>> The warning seems worth-while, even if the person receiving the message is
>> not in a position to do anything about it; maybe they'll hassle their sys
>> admin? Is there a better way of phrasing the message to make it more
>> informative / tolerable?
>>
>> Martin
>>
>>>
>>> Best,
>>> Kasper
>>>
>>>
>>>
>>>
>>> On Sun, Jun 1, 2014 at 8:19 AM, Wolfgang Huber <whuber at embl.de> wrote:
>>>
>>>
>>>> On 31 May 2014, at 23:53, Steve Lianoglou <lianoglou.steve at gene.com>
>>>> wrote:
>>>>
>>>>> On Sat, May 31, 2014 at 1:46 PM, Felix Francis [guest]
>>>>>
>>>>>> While trying to install Bioconductor, I get this warning message:
>>>>>> installed directory not writable, cannot update packages 'boot',
>>>>>>
>>>>> 'class', 'cluster', 'foreign', 'KernSmooth', 'lattice', 'MASS',
>>>>
>>>>>   'Matrix', 'mgcv', 'nlme', 'nnet', 'rpart', 'spatial'
>>>>>>
>>>>>
>>>>> It sounds like you do not have permissions to write into the installed
>>>>> directory that R was installed in on the computer you are using.
>>>>>
>>>>> Running the update/install as an administrator on the machine would
>>>>> likely do the trick.
>>>>>
>>>>
>>>> Depending on how valuable the machine is, or how many users work with it,
>>>> it is preferable to install R as well as its libraries with an account
>>>> with
>>>> limited rights; and then to use that account also for package updates and
>>>> the installation of additional packages, as what Felix is trying to do.
>>>>
>>>> Doing it as administrator seems like leaving a hole for through which bad
>>>> things could be done (intentionally or unintentionally) . Iâ EURO (tm)ve not yet
>>>> seen
>>>>
>>>> this actually happening, though.
>>>>
>>>>          Wolfgang
>>>>
>>>> _______________________________________________
>>>> Bioconductor mailing list
>>>> Bioconductor at r-project.org
>>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>>> Search the archives:
>>>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>>>
>>>>
>>>         [[alternative HTML version deleted]]
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Bioconductor mailing list
>>> Bioconductor at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>>> Search the archives: http://news.gmane.org/gmane.
>>> science.biology.informatics.conductor
>>>
>>>
>>
>> --
>> Computational Biology / Fred Hutchinson Cancer Research Center
>> 1100 Fairview Ave. N.
>> PO Box 19024 Seattle, WA 98109
>>
>> Location: Arnold Building M1 B861
>> Phone: (206) 667-2793
>>
>
>         [[alternative HTML version deleted]]
>
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor



-- 
Steve Lianoglou
Computational Biologist
Genentech