ATLAS Software - Revision history

Lhep at 13:19, 23 June 2016

2016-06-23T13:19:25Z

@@ Line 31: / Line 31: @@
-=== Use DQ2 ===
+=== Use Rucio ===
-To be able to use dq2 command, you need to first setup the grid access.
+To be able to use the rucio commands, you need to setup CVMFS first:
-Then, just make :
+    export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
-    source /afs/cern.ch/atlas/offline/external/GRID/ddm/DQ2Clients/setup.sh
+   source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh
-    export DQ2_LOCAL_SITE_ID=UNIBE-LHEP_LOCALGROUPDISK
+    alias asetup='source $AtlasSetup/scripts/asetup.sh'
-You able to search for datasets, look at the number of files in the dataset, etc... (follow this link[https://twiki.cern.ch/twiki/bin/viewauth/Atlas/DQ2ClientsHowTo] if you don't know the dq2 commands)
+Then type
+   lsetup rucio
 === Athena setup with CVMFS ===
-It is '''highly''' recommended to use CVMFS instead of AFS if using an Athena kit on one of the SLC5 machines in LHEP. CVMFS is a caching file system which hosts all athena releases and conditions data flat files. It was initially developed for the Cern-VM [https://twiki.cern.ch/twiki/bin/viewauth/Atlas/CernVM] Virtualization Software.
+CVMFS is a caching file system which hosts all athena releases and conditions data flat files. It was initially developed for the Cern-VM [https://twiki.cern.ch/twiki/bin/viewauth/Atlas/CernVM] Virtualization Software.
 <h4> Athena Setup </h4>
 To setup a particular release and testarea use the following commands and replace 17.0.2 and /path/to/my/testarea appropriately:
-  source /cvmfs/atlas.cern.ch/repo/sw/software/i686-slc5-gcc43-opt/17.0.2/cmtsite/asetup.sh 17.0.2
+   export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
-  asetup 17.0.2 --testarea /path/to/my/testarea
+   source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh
 <h4> Conditions DB files </h4>
@@ Line 60: / Line 55: @@
    /cvmfs/atlas-condb.cern.ch/repo/conditions/
@@ Line 79: / Line 68: @@
     cd ../run
     And finally run your code calling your program (the name defined in the cmt/requirement file)

Lhep at 13:14, 23 June 2016

2016-06-23T13:14:22Z

← Older revision		Revision as of 15:14, 23 June 2016
Line 1:		Line 1:
	=== Interactive machines at LHEP ===		=== Interactive machines at LHEP ===
	~~lheppc1~~.unibe.ch		ui01.lhep.unibe.ch
	~~lheppc7~~.unibe.ch		ui02.lhep.unibe.ch
			ui03.lhep.unibe.ch

	=== Basic SVN ===		=== Basic SVN ===
Line 12:		Line 12:

	=== LCG Grid Access ===		=== LCG Grid Access ===
	To setup the lcg grid environment, ~~type the following~~ :		To setup the lcg grid environment, setup CVMFS first :
	~~source~~ /~~terabig~~/~~atlsoft~~/~~LocalSoft~~/~~glite/glite-UI-current/external~~/~~etc/profile.d~~/~~grid-env~~.sh ~~&& unset GT_PROXY_MODE~~		export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
			source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh
			alias asetup='source $AtlasSetup/scripts/asetup.sh'
	=~~== Nordugrid Grid Access ===~~		Then type:
	~~To send jobs to Swiss clusters, you need to setup the nordugrid standalone client :~~		lsetup emi
	~~cd /terabig~~/~~atlsoft~~/~~LocalSoft/nordugrid-arc/nordugrid-arc-ui-current~~
	~~source setup~~.sh
	~~Beware 1~~: ~~You have to go into the nordugrid directory and then setup from this location, otherwise it does not work~~

	~~Beware 2: Both LCG and Nordugrid use the same instance of your grid certificate, so you don't have to change/renew your grid proxy, but you '''can not''' setup both environments in the same xterm.~~


	=== Prepare and check your grid certificate ===		=== Prepare and check your grid certificate ===
	To prepare your grid certificate, follow the following tutorial (you can adapt the procedure to our interactive nodes):		To prepare your grid certificate, follow the following tutorial (you can adapt the procedure to our interactive nodes):
	https://twiki.cern.ch/twiki/bin/~~viewauth~~/~~Atlas~~/~~FullGangaAtlasTutorial#3_1_Preparing_your_Grid_Certific~~		https://twiki.cern.ch/twiki/bin/view/AtlasComputing/WorkBook
			https://twiki.cern.ch/twiki/bin/view/AtlasComputing/WorkBookStartingGrid (Preparing your certificate)

	To verify the expiry date of your grid certificate installed in the .globus subdir of your home directory (on teh interactive nodes and/or lxplus):		To verify the expiry date of your grid certificate installed in the .globus subdir of your home directory (on teh interactive nodes and/or lxplus):

Lhep: 1 revision imported

2015-03-18T07:49:23Z

1 revision imported

← Older revision	Revision as of 09:49, 18 March 2015
(No difference)

Lhep at 10:57, 12 December 2011

2011-12-12T10:57:44Z

New page

=== Interactive machines at LHEP ===
lheppc1.unibe.ch
lheppc7.unibe.ch

=== Basic SVN ===
svn mkdir -m “Add OQMaps dir" http://svn.lhep.unibe.ch/subv/atlas/SUSY2011/cutflow/OQMaps
svn co http://svn.lhep.unibe.ch/subv/atlas/SUSY2011/cutflow/OQMaps
svn add * # All files you have in OQMaps. Directories must be created like above.
svn ci "Add OQMaps”

=== LCG Grid Access ===
To setup the lcg grid environment, type the following :
source /terabig/atlsoft/LocalSoft/glite/glite-UI-current/external/etc/profile.d/grid-env.sh && unset GT_PROXY_MODE

=== Nordugrid Grid Access ===
To send jobs to Swiss clusters, you need to setup the nordugrid standalone client :
cd /terabig/atlsoft/LocalSoft/nordugrid-arc/nordugrid-arc-ui-current
source setup.sh
Beware 1: You have to go into the nordugrid directory and then setup from this location, otherwise it does not work

Beware 2: Both LCG and Nordugrid use the same instance of your grid certificate, so you don't have to change/renew your grid proxy, but you '''can not''' setup both environments in the same xterm.

=== Prepare and check your grid certificate ===
To prepare your grid certificate, follow the following tutorial (you can adapt the procedure to our interactive nodes):
https://twiki.cern.ch/twiki/bin/viewauth/Atlas/FullGangaAtlasTutorial#3_1_Preparing_your_Grid_Certific

To verify the expiry date of your grid certificate installed in the .globus subdir of your home directory (on teh interactive nodes and/or lxplus):
cd ~/.globus
openssl x509 -in usercert.pem -noout -enddate
Follow this link[http://security.ncsa.illinois.edu/research/grid-howtos/usefulopenssl.html] for a summary of useful openssl commands.

=== Use DQ2 ===
To be able to use dq2 command, you need to first setup the grid access.
Then, just make :
source /afs/cern.ch/atlas/offline/external/GRID/ddm/DQ2Clients/setup.sh
export DQ2_LOCAL_SITE_ID=UNIBE-LHEP_LOCALGROUPDISK
You able to search for datasets, look at the number of files in the dataset, etc... (follow this link[https://twiki.cern.ch/twiki/bin/viewauth/Atlas/DQ2ClientsHowTo] if you don't know the dq2 commands)

'''But''', if you want to '''download locally a dataset''', you have also to setup the Athena environment to fix the problem with python releases between dq2 and the grid environment. You can setup the Athena environment just before or after the dq2 setup, as you want.

'''Not sure the above is still relevant...'''

=== Athena setup with CVMFS ===
It is '''highly''' recommended to use CVMFS instead of AFS if using an Athena kit on one of the SLC5 machines in LHEP. CVMFS is a caching file system which hosts all athena releases and conditions data flat files. It was initially developed for the Cern-VM [https://twiki.cern.ch/twiki/bin/viewauth/Atlas/CernVM] Virtualization Software.

<h4> Athena Setup </h4>
To setup a particular release and testarea use the following commands and replace 17.0.2 and /path/to/my/testarea appropriately:
source /cvmfs/atlas.cern.ch/repo/sw/software/i686-slc5-gcc43-opt/17.0.2/cmtsite/asetup.sh 17.0.2
asetup 17.0.2 --testarea /path/to/my/testarea

<h4> Alternative Athena Setup (suitable for start-up scripts, e.g. .bash_profile)</h4>
export ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
source ${ATLAS_LOCAL_ROOT_BASE}/user/atlasLocalSetup.sh
asetup 17.0.2 --single --testarea /path/to/my/testarea

<h4> Conditions DB files </h4>
The conditions DB flat files can be found in:
/cvmfs/atlas-condb.cern.ch/repo/conditions/

<h4> Development Nightly Builds (64 bit) </h4>
source /cvmfs/atlas-nightlies.cern.ch/repo/sw/nightlies/x86_64-slc5-gcc43-opt/17.X.0/rel_4/cmtsite/setup.sh -tag=AtlasProduction,rel_4,opt,gcc43,slc5,64
asetup AtlasProduction,rel_4,64 --gccversion 4.3.5 --gcclocation /cvmfs/atlas-nightlies.cern.ch/repo/sw/nightlies/x86_64-slc5-gcc43-opt/17.X.0/atlas-gcc-435/slc5/

<h4> Bugfix Nightly Builds </h4>
source /cvmfs/atlas-nightlies.cern.ch/repo/sw/nightlies/i686-slc5-gcc43-opt/17.0.X/rel_4/cmtsite/setup.sh -tag=AtlasProduction,rel_4,opt,gcc43,slc5,32

=== How to use ROOT libraries from gcc ===
CINT, the compiler included within ROOT, is not fully reliable from my point of view just for at least two reasons. First, it allows some "bad code writing" hard to debug once you change your ROOT release. Second, CINT, as python, is not a real compiler but just interpret your code for each event. So, if your code becomes sophisticated (kinematic fits,...) CINT spend much more time to interpret your code for each event than running on the event. But, you can use ROOT as libraries from gcc, in a CMT environment in the following :

source /terabig/atlsoft/LocalSoft/tools/config/setup.sh

Then, go in your package/cmt directory (an example can be found here [http://svn.lhep.unibe.ch/subv/users/cogneras/ROOTAnalysis/]) This package, written for the SUSY/Trigger analysis as been documented using Doxygen. The documentation can be found here[http://cogneras.web.cern.ch/cogneras/restricted/Doxy_SusyAnalysis/index.html](page reachable only from unibe.ch domain)
cmt config
source setup.sh
gmake (or make)
cd ../run
And finally run your code calling your program (the name defined in the cmt/requirement file)

=== Obsolete pages ===
<h4> Version 13 Bern ATLAS Cluster - SLC4 (OBSOLETE) </h4>

* ATLAS Software 13.0.x uses a root version, which wants to get the home directory of the user. On lheppc* ldap setup, nssswitch (I think) calls the nss_ldap library, which is not installed on standard SLC4. root crashes because of this. Need to install (yum install nss_ldap) on SLC4 nodes for LHEP cluster.
* MooEventCompilation needs crti.o, which is provides by package glibc-devel.i386. So need to install this
* There are a few symlinks missing, which cause MooEventCompile to crash. Thanks to szymon for the fix.

#as root, execute the following lines to make ATLAS software release 13 work
yum install nss_ldap
yum install glibc-devel.i386
cd /usr/lib
ln -s /usr/lib/libjpeg.so.62 libjpeg.so
ln -s /usr/lib/libfreetype.so.6 libfreetype.so
ln -s /usr/lib/libg2c.so.0 libg2c.so
cd /usr/X11R6/lib/
ln -s /usr/X11R6/lib/libXm.so.3 libXm.so

If you want to automatize the process, copy this code in a file fixnode.sh, and execute as root

yes | sh fixnode.sh

<h4> Version 14 Bern ATLAS Cluster - SLC4 (OBSOLETE) </h4>

* ATLAS Software 14.x needs a few packages. They are listed in /ngsys/slcrequiredpackages/reqpack.txt
#to install, run as root:
cd /ngsys/slcrequiredpackages
sh install.sh

<h4> Version 13 UBELIX Cluster - Gentoo Linux (OBSOLETE) </h4>

Changes with every release. Follow runtime environment script to see al the tweaks made. (/home/ubelix/lhep/nordugrid/software/APPS/HEP)