Network install of packages
1
0
Entering edit mode
@anthony-mullen-3127
Last seen 8.4 years ago
Greetings - I am a Linux system administrator for the Department of Mathematics and the University of Illinois. I installed R, perl-bioperl and perl-bioperl-run for a couple researchers who are doing collaboration using these tools. I was asked to install Bioconductor packages. I took a quick look at the installation notes at the bioconductor.org page. And I have a few questions. Where can I find the sizes of these standard/additional packages? If I install these packages as root will they be available to other users? Or will each person have to install Bioconductor packages themselves? I think I see what I need to do if I were to install these packages. But if Bioconductor packages are installed only in a local user-space. Then it doesn't make any sense for me to install them as root. It looks like the commands I may want to use are: #invoke R source("http://bioconductor.org/biocLite.R") biocLite() biocLite(groupName="all") But I see a note that states that selecting 'all' software packages will take up a lot of space. But if the packages can be installed by root, such that other's can use the packages. I can install all the Bioconductor packages on a local computer. And the users network home directories won't have to support all the disk space. I hope this all makes sense. If you need any more clarification please don't hesitate in letting me know. Tony -
Network Network • 1.2k views
ADD COMMENT
0
Entering edit mode
@sean-davis-490
Last seen 5 months ago
United States
On Thu, Nov 6, 2008 at 1:11 PM, Anthony Mullen <mullen1@uiuc.edu> wrote: > Greetings - > > I am a Linux system administrator for the Department of Mathematics and the > University of Illinois. I installed R, perl-bioperl and perl- bioperl-run > for a couple researchers who are doing collaboration using these tools. > > I was asked to install Bioconductor packages. I took a quick look at the > installation notes at the bioconductor.org page. And I have a few > questions. > > Where can I find the sizes of these standard/additional packages? If I > install these packages as root will they be available to other users? Or > will each person have to install Bioconductor packages themselves? > > I think I see what I need to do if I were to install these packages. But if > Bioconductor packages are installed only in a local user-space. Then it > doesn't make any sense for me to install them as root. > > It looks like the commands I may want to use are: > > #invoke R > source("http://bioconductor.org/biocLite.R") > biocLite() > biocLite(groupName="all") > > But I see a note that states that selecting 'all' software packages will > take up a lot of space. But if the packages can be installed by root, such > that other's can use the packages. I can install all the Bioconductor > packages on a local computer. And the users network home directories won't > have to support all the disk space. > Hi, Tony. This is not a bioconductor issue, directly, since R deals with this issue out-of-the-box. If you install as root, then the users will have access to those packages. If they need other packages, the packages will, by default, be installed in the user's home folder. So, you can use exactly the commands you mention above to install bioconductor and then, if further packages are necessary, the user can use biocLite() (and I would HIGHLY RECOMMEND using biocLite() for all R package installation, if possible) to install those other packages and R will "do the right thing" and install them into the home directory. Sean [[alternative HTML version deleted]]
ADD COMMENT
0
Entering edit mode
Sean - > Hi, Tony. This is not a bioconductor issue, directly, since R deals > with this issue out-of-the-box. If you install as root, then the users > will have access to those packages. Excellent. That's what I needed to know. We have a large network share. And use NFS to export home directories as a user logs in to a computer. So basically all of our Linux computers have a lot of space on their disks for additional packages. And I would much rather install to a local computer, and use it's under-utilized disk space. Rather than a couple people grab their own copies of software. ;-) > If they need other packages, [...] > use biocLite() (and I would HIGHLY RECOMMEND using biocLite() for all R > package installation, if possible) [...] I'll just install everything I can. And then the researchers can sort things out. ;-) [... time goes by ...] And after running the aforementioned commands to install all the Bioconductor packages I have a lot of gziped tar-files in a temporary directory (/tmp/RtmpJAenmw/downloaded_packages). And a lot of similarly-named files in /usr/lib/R/library. So the installation of Bioconductor was completely painless. Thank you for your time. And the little bit of hand-holding that I needed. ;-) Tony -
ADD REPLY
0
Entering edit mode
Hi Anthony, I think it would be good sys-ad practice to install R and its packages (whether or not they're from Bioconductor) under a user-id that has been allowed to write to a shared directory, but that has otherwise restricted rights - but *not as root*. Best wishes# Wolfgang PS Like your. Creative. Use of Punctation. ---------------------------------------------------- Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber Sean Davis ha scritto: > On Thu, Nov 6, 2008 at 1:11 PM, Anthony Mullen <mullen1 at="" uiuc.edu=""> wrote: > >> Greetings - >> >> I am a Linux system administrator for the Department of Mathematics and the >> University of Illinois. I installed R, perl-bioperl and perl- bioperl-run >> for a couple researchers who are doing collaboration using these tools. >> >> I was asked to install Bioconductor packages. I took a quick look at the >> installation notes at the bioconductor.org page. And I have a few >> questions. >> >> Where can I find the sizes of these standard/additional packages? If I >> install these packages as root will they be available to other users? Or >> will each person have to install Bioconductor packages themselves? >> > >> I think I see what I need to do if I were to install these packages. But if >> Bioconductor packages are installed only in a local user-space. Then it >> doesn't make any sense for me to install them as root. >> >> It looks like the commands I may want to use are: >> >> #invoke R >> source("http://bioconductor.org/biocLite.R") >> biocLite() >> biocLite(groupName="all") >> >> But I see a note that states that selecting 'all' software packages will >> take up a lot of space. But if the packages can be installed by root, such >> that other's can use the packages. I can install all the Bioconductor >> packages on a local computer. And the users network home directories won't >> have to support all the disk space. >> > > > Hi, Tony. This is not a bioconductor issue, directly, since R deals with > this issue out-of-the-box. If you install as root, then the users will have > access to those packages. If they need other packages, the packages will, > by default, be installed in the user's home folder. So, you can use exactly > the commands you mention above to install bioconductor and then, if further > packages are necessary, the user can use biocLite() (and I would HIGHLY > RECOMMEND using biocLite() for all R package installation, if possible) to > install those other packages and R will "do the right thing" and install > them into the home directory. > > Sean > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
ADD REPLY
0
Entering edit mode
Has anyone done any benchmarking on accessing the annotation packages over a network vs. a local drive? I can image (guess) that for some annotation packages that are frequently used, it may pay of to have the on a local drive, especially the SQLite based ones. Correct me if I'm wrong. /Henrik On Thu, Nov 6, 2008 at 12:00 PM, Wolfgang Huber <huber at="" ebi.ac.uk=""> wrote: > > Hi Anthony, > > I think it would be good sys-ad practice to install R and its packages > (whether or not they're from Bioconductor) under a user-id that has been > allowed to write to a shared directory, but that has otherwise restricted > rights - but *not as root*. > > Best wishes# > Wolfgang > > PS Like your. Creative. Use of Punctation. > > ---------------------------------------------------- > Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber > > > Sean Davis ha scritto: >> >> On Thu, Nov 6, 2008 at 1:11 PM, Anthony Mullen <mullen1 at="" uiuc.edu=""> wrote: >> >>> Greetings - >>> >>> I am a Linux system administrator for the Department of Mathematics and >>> the >>> University of Illinois. I installed R, perl-bioperl and perl- bioperl-run >>> for a couple researchers who are doing collaboration using these tools. >>> >>> I was asked to install Bioconductor packages. I took a quick look at the >>> installation notes at the bioconductor.org page. And I have a few >>> questions. >>> >>> Where can I find the sizes of these standard/additional packages? If I >>> install these packages as root will they be available to other users? Or >>> will each person have to install Bioconductor packages themselves? >>> >> >>> I think I see what I need to do if I were to install these packages. But >>> if >>> Bioconductor packages are installed only in a local user-space. Then it >>> doesn't make any sense for me to install them as root. >>> >>> It looks like the commands I may want to use are: >>> >>> #invoke R >>> source("http://bioconductor.org/biocLite.R") >>> biocLite() >>> biocLite(groupName="all") >>> >>> But I see a note that states that selecting 'all' software packages will >>> take up a lot of space. But if the packages can be installed by root, >>> such >>> that other's can use the packages. I can install all the Bioconductor >>> packages on a local computer. And the users network home directories >>> won't >>> have to support all the disk space. >>> >> >> >> Hi, Tony. This is not a bioconductor issue, directly, since R deals with >> this issue out-of-the-box. If you install as root, then the users will >> have >> access to those packages. If they need other packages, the packages will, >> by default, be installed in the user's home folder. So, you can use >> exactly >> the commands you mention above to install bioconductor and then, if >> further >> packages are necessary, the user can use biocLite() (and I would HIGHLY >> RECOMMEND using biocLite() for all R package installation, if possible) to >> install those other packages and R will "do the right thing" and install >> them into the home directory. >> >> Sean >> >> [[alternative HTML version deleted]] >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at stat.math.ethz.ch >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor > > _______________________________________________ > Bioconductor mailing list > Bioconductor at stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor >
ADD REPLY
0
Entering edit mode
On Thu, Nov 6, 2008 at 3:05 PM, Henrik Bengtsson <hb@stat.berkeley.edu>wrote: > Has anyone done any benchmarking on accessing the annotation packages > over a network vs. a local drive? I can image (guess) that for some > annotation packages that are frequently used, it may pay of to have > the on a local drive, especially the SQLite based ones. Correct me if > I'm wrong. > The hit will depend on numerous hardware and computer environment factors, particularly those which affect the ability to keep the SQLite database in memory (cached). Sean > > /Henrik > > On Thu, Nov 6, 2008 at 12:00 PM, Wolfgang Huber <huber@ebi.ac.uk> wrote: > > > > Hi Anthony, > > > > I think it would be good sys-ad practice to install R and its packages > > (whether or not they're from Bioconductor) under a user-id that has been > > allowed to write to a shared directory, but that has otherwise restricted > > rights - but *not as root*. > > > > Best wishes# > > Wolfgang > > > > PS Like your. Creative. Use of Punctation. > > > > ---------------------------------------------------- > > Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber > > > > > > Sean Davis ha scritto: > >> > >> On Thu, Nov 6, 2008 at 1:11 PM, Anthony Mullen <mullen1@uiuc.edu> > wrote: > >> > >>> Greetings - > >>> > >>> I am a Linux system administrator for the Department of Mathematics and > >>> the > >>> University of Illinois. I installed R, perl-bioperl and > perl-bioperl-run > >>> for a couple researchers who are doing collaboration using these tools. > >>> > >>> I was asked to install Bioconductor packages. I took a quick look at > the > >>> installation notes at the bioconductor.org page. And I have a few > >>> questions. > >>> > >>> Where can I find the sizes of these standard/additional packages? If I > >>> install these packages as root will they be available to other users? > Or > >>> will each person have to install Bioconductor packages themselves? > >>> > >> > >>> I think I see what I need to do if I were to install these packages. > But > >>> if > >>> Bioconductor packages are installed only in a local user-space. Then it > >>> doesn't make any sense for me to install them as root. > >>> > >>> It looks like the commands I may want to use are: > >>> > >>> #invoke R > >>> source("http://bioconductor.org/biocLite.R") > >>> biocLite() > >>> biocLite(groupName="all") > >>> > >>> But I see a note that states that selecting 'all' software packages > will > >>> take up a lot of space. But if the packages can be installed by root, > >>> such > >>> that other's can use the packages. I can install all the Bioconductor > >>> packages on a local computer. And the users network home directories > >>> won't > >>> have to support all the disk space. > >>> > >> > >> > >> Hi, Tony. This is not a bioconductor issue, directly, since R deals > with > >> this issue out-of-the-box. If you install as root, then the users will > >> have > >> access to those packages. If they need other packages, the packages > will, > >> by default, be installed in the user's home folder. So, you can use > >> exactly > >> the commands you mention above to install bioconductor and then, if > >> further > >> packages are necessary, the user can use biocLite() (and I would HIGHLY > >> RECOMMEND using biocLite() for all R package installation, if possible) > to > >> install those other packages and R will "do the right thing" and install > >> them into the home directory. > >> > >> Sean > >> > >> [[alternative HTML version deleted]] > >> > >> _______________________________________________ > >> Bioconductor mailing list > >> Bioconductor@stat.math.ethz.ch > >> https://stat.ethz.ch/mailman/listinfo/bioconductor > >> Search the archives: > >> http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > _______________________________________________ > > Bioconductor mailing list > > Bioconductor@stat.math.ethz.ch > > https://stat.ethz.ch/mailman/listinfo/bioconductor > > Search the archives: > > http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]
ADD REPLY
0
Entering edit mode
> Has anyone done any benchmarking on accessing the annotation packages > over a network vs. a local drive? I can image (guess) that for some > annotation packages that are frequently used, it may pay of to have > the on a local drive, especially the SQLite based ones. Correct me if > I'm wrong. Using AFS as your network file system reduces the impact of this issue or may eliminate it completely - as whatever you've read from the network is cached on local disk. The original poster has the admirable aim of utilizing the under-used local disk space on a large number of boxes. However, I'd still vote for keeping a single copy of everything rather than n copies. -- Atro Tossavainen (Mr.) / The Institute of Biotechnology at Systems Analyst, Techno-Amish & / the University of Helsinki, Finland, +358-9-19158939 UNIX Dinosaur / employs me, but my opinions are my own. < URL : http : / / www . helsinki . fi / %7E atossava / > NO FILE ATTACHMENTS
ADD REPLY

Login before adding your answer.

Traffic: 234 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6