Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.python > #97978 > unrolled thread

Is it ok to install python binaries on a network drive?

Back to article view | Back to comp.lang.python

Contents

  Is it ok to install python binaries on a network drive? Fabien <fabien.maussion@gmail.com> - 2015-10-28 13:11 +0100
    Re: Is it ok to install python binaries on a network drive? harirammanohar159@gmail.com - 2015-10-29 01:44 -0700
      Re: Is it ok to install python binaries on a network drive? Fabien <fabien.maussion@gmail.com> - 2015-10-29 11:42 +0100
    Re: Is it ok to install python binaries on a network drive? Steven D'Aprano <steve@pearwood.info> - 2015-10-29 23:07 +1100
      Re: Is it ok to install python binaries on a network drive? Fabien <fabien.maussion@gmail.com> - 2015-11-04 09:39 +0100

#97978 — Is it ok to install python binaries on a network drive?

Folks,

for a course I am teaching I thought about following solution to get 
everybody have the same environment: I installed python 3.4 and the 
packages I need (mostly SciPy stuffs) with miniconda (a scientific 
version of pip + virtualenv merged together) on a network folder. The 
students have r+x access to it.

The students will log in on their local machine, add the network folder 
(e.g. /scratch/fabiens_folder/bin/) to their $PATH and everything seems 
to work (I just tried it).

The first "import numpy" takes a bit more time than when the package is 
installed locally, but after the first import it seems to work as expected.

My question is: what will happen when about 20 students will do the same 
at the same time? I know too little about what's going under the hood 
with an "import X" command to be able to predict problems that might occur.

If that matters: the network storage where python is installed and the 
place where the course take place are located in the same university, so 
that the network communication *should* be quite fast.

Thanks for your thoughts,

Fabie

[toc] | [next] | [standalone]

#97992

On Wednesday, 28 October 2015 17:41:49 UTC+5:30, Fabien  wrote:
> Folks,
> 
> for a course I am teaching I thought about following solution to get 
> everybody have the same environment: I installed python 3.4 and the 
> packages I need (mostly SciPy stuffs) with miniconda (a scientific 
> version of pip + virtualenv merged together) on a network folder. The 
> students have r+x access to it.
> 
> The students will log in on their local machine, add the network folder 
> (e.g. /scratch/fabiens_folder/bin/) to their $PATH and everything seems 
> to work (I just tried it).
> 
> The first "import numpy" takes a bit more time than when the package is 
> installed locally, but after the first import it seems to work as expected.
> 
> My question is: what will happen when about 20 students will do the same 
> at the same time? I know too little about what's going under the hood 
> with an "import X" command to be able to predict problems that might occur.
> 
> If that matters: the network storage where python is installed and the 
> place where the course take place are located in the same university, so 
> that the network communication *should* be quite fast.
> 
> Thanks for your thoughts,
> 
> Fabie

Actually you can do one thing, instruct your students to download the package to local and start programming for a session, so that updated package has to be available on every ones desktop....

[toc] | [prev] | [next] | [standalone]

#97995

On 10/29/2015 09:44 AM, harirammanohar159@gmail.com wrote:
> Actually you can do one thing, instruct your students
 > to download the package to local

I wish I could: they have approx 1Gb space on their $home and miniconda 
+ packages alone needs that space...

Thanks!

Fabien

[toc] | [prev] | [next] | [standalone]

#97997

On Wed, 28 Oct 2015 11:11 pm, Fabien wrote:

> Folks,
> 
> for a course I am teaching I thought about following solution to get
> everybody have the same environment: I installed python 3.4 and the
> packages I need (mostly SciPy stuffs) with miniconda (a scientific
> version of pip + virtualenv merged together) on a network folder. The
> students have r+x access to it.
> 
> The students will log in on their local machine, add the network folder
> (e.g. /scratch/fabiens_folder/bin/) to their $PATH and everything seems
> to work (I just tried it).
> 
> The first "import numpy" takes a bit more time than when the package is
> installed locally, but after the first import it seems to work as
> expected.
> 
> My question is: what will happen when about 20 students will do the same
> at the same time? I know too little about what's going under the hood
> with an "import X" command to be able to predict problems that might
> occur.

I think it should be fine. I don't expect there to be any problems: your
students don't have write access to the shared area. Applications on a
network drive may be ever-so-slightly slower than local applications, but I
don't imagine it will be too bad. Depending on the network, there may be a
noticeable lag when all 20 students try to launch Python at the same time,
but once they're actually running it, it should be fine.

Possibly a slightly better approach might be to install Python and the
standard library on each individual workstation/desktop/laptop, install the
extra packages on the network drive, and then add that network drive to the
PYTHONPATH of each workstation. (Use a .pth file in their site-packages
directory.)




-- 
Steven

[toc] | [prev] | [next] | [standalone]

#98213

On 10/29/2015 01:07 PM, Steven D'Aprano wrote:
>> My question is: what will happen when about 20 students will do the same
>> >at the same time? I know too little about what's going under the hood
>> >with an "import X" command to be able to predict problems that might
>> >occur.
> I think it should be fine.

Dear all,

I just wanted to report that everything worked fine. It wasn't even too 
slow. I find it a quite elegant solution after all for all the students 
to have the same working environment.

Thanks!

Fabien

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.python

csiph-web