1 / 7

How to “use” CMSSW on own Linux Box and be happy

How to “use” CMSSW on own Linux Box and be happy. In this context “use” means: check-out pre-compiled CMSSW code run pre-compiled code compile and then run my modified code commit to the central CVS repository my code

gudrun
Télécharger la présentation

How to “use” CMSSW on own Linux Box and be happy

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. How to “use” CMSSW on own Linux Box and be happy • In this context “use” means: • check-out pre-compiled CMSSW code • run pre-compiled code • compile and then run my modified code • commit to the central CVS repository my code • do all the previous things on top of a CMSSW (pre)-release of my choice Own linux box means: - the desktop in my office or even my laptop. Something I can access directly. No ssh to lxplus, uaf.t2.ucsd, uaf.fnal,etc..

  2. Motivations • Most of us has a 64-bit dual (quad) core machine which is used just to log into another machine where the job is actually run • uaf-1.ucsd.edu ( the currently most used machine to actually run CMSSW job) will not be enough in a short while. The number of real CMSSW users at UCSD is increasing • To develop code, often special release or nightly build are necessary. FNAL and Cern provides machine where all such releaseS are installed, but the connection is slow and painful. I want to log into FNAL only for mass production jobs! • I want to run Iguana (at least try...they told me that it is much faster now). And it can be “decently” used only from a local machine: Even if Iguana was speed-light fast, forwarding its graphic output through a ssh connection makes the user job much slow. No hope.

  3. First installation of CMSSW • the installation of the first CMSSW release is lengthy because all the external packages CMSSW depends on (root, clhep, geant, special version of gcc,etc..) have to be installed too. It is about 2 GB of stuff. • The most complete guide I found is here: https://twiki.cern.ch/twiki/bin/view/CMS/CMSSW_aptinstaller • If you plan to compile your own CMSSW code, special care has to be taken on 64-bit machine because CMSSW has to be compiled in 32-bit compatibility mode. An additional 32 bit library has to be installed: yum install glibc-devel-2.3.4-2.36.x86-32.rpm By default only glibc-devel-2.3.4-2.36.x86-64.rpm is installed

  4. 2nd and further installations of CMSSW • It is pretty quick and easy. It takes no more than 15 minutes: • without external package (already installed), a fresh CMSSW release is no more than 150 MB. export VO_CMS_SW_DIR=/code/osgcode/cmslocal/cms eval `$VO_CMS_SW_DIR/aptinstaller.sh config -path $VO_CMS_SW_DIR -arch slc4_ia32_gcc345 -csh` apt-get update apt-get install cms+cmssw+CMSSW_1_5_0_pre4 directory of your choice used during the first installation here you choose the (pre-)release you want

  5. check-out and commit code from/to the central repository To just check-out code from the central CMSSW repository, the standard known recipe is enough: export CVSROOT=:pserver:anonymous@cmscvs.cern.ch:/cvs_server/repositories/CMSSW 2) cvs login (use ‘98passwd’ as password) 3),4),... cvs co -rCMSSW_X_Y_Z SubSystem/Package To commit code, you need to authenticate yourself. There are 2 options. A) authenticate yourself every time you use a cvs command: export CVSROOT=:kserver:cmscvs.cern.ch:/cvs_server/repositories/CMSSW 2),3),4),... cvs -d :ext:yourName@cmscvs.cern.ch:/cvs_server/repositories/CMSSW commit (insert your password for your ‘yourName’ account at cern)

  6. check-out and commit code from/to the central repository B) authenticate yourself once (using a kerberos ticket): 1) export CVSROOT=:kserver:cmscvs.cern.ch:/cvs_server/repositories/CMSSW 2) kinit -4 -l 2d yourName@CERN.CH (insert your password only once) 3),4),... cvs commit (no password is requested this time) Nevertheless, to make kerberos4 works, these lines as to be added to 2 configuration files: to be done only once. These line can go in the .bashrc file CERN.CH CERN.CH afsdb1.cern.ch CERN.CH afsdb3.cern.ch CERN.CH afsdb2.cern.ch /etc/krb.conf /etc/krb.realms .cern.ch CERN.CH

  7. summary • Milestone: • Yesterday I succeeded to install the last pre-release available (150pre4), fix some code, test it and commit it back to CVS. All from my desktop, all in few minutes • Next: • try to use Iguana locally and see what happens. Thanks to J.Letts and O.Gutsche for the help

More Related