[Bioc-devel] Announcing Docker containers for Bioconductor

Elena Grassi grassi.e at gmail.com
Thu Jan 8 17:45:37 CET 2015


On Thu, Jan 8, 2015 at 4:43 PM, Dan Tenenbaum <dtenenba at fredhutch.org> wrote:
>> I have a strange (ok, at least to me :) ) issue concerning volumes.
>> On a machine (debian testing/unstable up to date) everything works
>> smoothly when I do something like:
>> data at decoder:~$ docker run -v
>> /home/data/Dropbox/work/
>> matrix/mr_bioc/matrix_rider/:/opt/matrix_rider
>> -p 8787:8787 9f40f8036ad4
>> (in /home/data/Dropbox/work/matrix/mr_bioc/matrix_rider/ I have a
>> Rstudio project)
>
> What image is 9f40f8036ad4? Why are you not using the user/repository name for the image? Is it a rocker image or a Bioconductor image?

Sorry! It's bioconductor/devel_sequencing image that I had pulled before:
data at decoder:~/Dropbox/work/matrix/matrix_rider$ docker images
REPOSITORY                      TAG                 IMAGE ID
 CREATED             VIRTUAL SIZE
bioconductor/devel_sequencing   latest              9f40f8036ad4
 29 hours ago        5.715 GB

> Note that Rstudio Server runs as a user called rstudio, not necessarily a privileged user.
> So the directory you are trying to mount should probably be writable (and readable) by all.
> so try
>
> chmod -R a+rw /home/data/Dropbox/work/matrix/mr_bioc/matrix_rider/

On the second machine this works but on this machine I do not love
having this kind of privileges as long as it's shared PC in my office
:)
What puzzles me it's that the privileges were (i.e. not writable and
readable by everyone) the same on both machines and on the first one
it has worked from the first moment without a+rw...I assumed that the
rstudio user has PID 1000 in the container/image and that -e UID= etc
could fix (change it to whatever PID my user has on the host machine)
this but this does not seem the case.

Thank you very much,
E.
-- 
$ pom



More information about the Bioc-devel mailing list