The SSM environment
The "SSM environment"
The "SSM environment" is build on SSM, Simple Software Management.
This means one now subscribes to domaines and packages using s.ssmuse.dot to aquire access to certain directories, scripts, binaries, etc. In general the s.ssmuse.dot will add directories to your $PATH, $LD_LIBRARY_PATH, etc and set some environment variables.
For interactive sessions I find the following packages very useful:
diagtools, difftools, runtools, vistools, fulldev
But first you will have to do the little "setup" below, then you can subscribe to one or more of these packages by typing for example:
. s.ssmuse.dot diagtools difftools runtools vistools fulldev
I added this line to my .interactive_profile (see below).
Setup the SSM environment
To use the SSM environment you will first (and only once) have to do a little "setup":
- Create two links in your HOME:
cd
ln -fs ~winger/ssm-domains/setup/v_001 ECssm
ln -fs ECssm/dot-profile-setup_1.0_multi/bin/.profile
Change the name of two files (or remove them):
mv .bash_profile .bash_profile-
mv .bashrc .bashrc-
Log out and back in. - There is no .profile_usr anymore. It got replaced by several .something_profile files which all have to reside in your HOME in a directory called '.profile.d'.
Have a look at my .profile.d:
ls -al ~winger/.profile.d
I suggest you link to the same group profiles I link to:
cd ~/.profile.d
ln -s ~winger/.published_profile.d/.group_profile
ln -s ~winger/.published_profile.d/.group0_profile
This will give you some basic aliases I find quite useful. Have a look at them if you like.
Your own personal aliases, exports, s.ssmuse.dot etc. for interactive use will go into the file
.interactive_profile
which you will have to create.
Then you either log out and back in again or source the new .profiles. - Allow 'ssh $TRUE_HOST' without typing password:
cd ~/.ssh
Only if there is no(!) file 'id_rsa.pub' create it with:
ssh-keygen (press just 'Enter' whenever asked a question, 3 times)
then
cat id_rsa.pub >> authorized_keys2 - Create directories for running CRCM5.
Create directory/link for the execution directory (where the model will be running):
mkdir ~/MODEL_EXEC_RUN
cd ~/MODEL_EXEC_RUN
mkdir -p /sb/project/xgk-345-ab/${USER}/EXECDIR
ln -s /sb/project/xgk-345-ab/${USER}/EXECDIR localhost
ln -s /sb/project/xgk-345-ab/${USER}/EXECDIR guillimin
Create directory/link for the listings:
mkdir ~/listings
cd ~/listings
mkdir -p /sb/project/xgk-345-ab/${USER}/Listings
ln -s /sb/project/xgk-345-ab/${USER}/Listings localhost
ln -s /sb/project/xgk-345-ab/${USER}/Listings guillimin - Set project name of which to use the core allocation
You will have to create a file in your HOME on guillimin which is called:
${HOME}/.CLUMEQ_accnt
This file has to contain 1 variable called 'RAPid', set to the RAPid of the project under which you want to submit your job.
For people working for or having their account via Pierre Gauthier, Laxmi Sushama, René Laprise or Peter Yau you will have to set in the above file:
RAPid=xgk-345-ac
Have a look at my file if you like:
~winger/.CLUMEQ_accnt - Open your HOME to the world
If you do want my or other help I suggest you open your HOME to the world:
chmod 755 /home/$USER