Getting started at LSCE (by JYP)

A few useful links. Some have nothing to do with science and LSCE hardware/software resources!

• LSCE intranet

• LSCE-related public transport
• Adding money online to your restaurant account
• Survival kit and other information for CDD and interns (LSCE intranet)
• International welcome desk Université Paris-Saclay supports international students, PhD students and researchers as soon as they arrive in France to ensure their stay take place under the best conditions
• Trouver un logement dans une résidence universitaire (in French… Finding a room in a student dorm)
• Science Accueil
  • Service request page
• Guide du Personnel UVSQ

• Phone directory (intranet)
  • Dial 2 + NNNN (last 4 digits) to call somebody at LSCE (or CEA Saclay) from LSCE
  • Dial 0 + NNNNNNNNNNNNNNNNNNNN to call somebody outside LSCE from LSCE
  • Dial 18 to call FLS (CEA security & firemen)
    • Other security related phone numbers
      • Recommended: add the FLS full phone number to your cell phone
• 701 and 714 map (intranet)

Use the following sequence of keys on the phone in your office:

1. Select MENU (then use the downward arrow)
2. Select PARAMETRES (use the downward arrow)
3. Select GENERAL (use the downward arrow)
4. Select MON NUMERO
   • This will display your 2 + NNNN phone number
   • You can be called from outside with 01 69 08 NN NN

1. Find your name in the intranet phone directory
2. Click on Modifier mes coordonnées (in blue on the right side)
3. Click on Informations dans l'annuaire and update the different fields
4. Click on Envoyez votre demande
5. Wait (for someone to validate your request)

It is important to join the correct mailing lists, if you want to get all the information you need, on time.

When your account is created, your email address should automatically be added to the the all list, that can be used to send mails to all the people working at LSCE. You will also get IPSL information through this list

Mails sent to these lists will reach lots of people. Think well before sending a mail to these lists!

Ask your team leader to add you to your team's mailing list. Example below: the CLIM and ESTIMR lists

• You have to be identified on the server to access most lists's details. Enter you login and email password in the email address and pasword fields at the top right of any mailing list web page
• You may want/need to join other mailing lists (project specific, sports, …). Check the list of lists to find about all the existing lists
• It is possible to use mails to interact with the LSCE lists server, rather than the web interface

You can't connect your personal laptop to the LSCE wired network (aka ethernet), it will not work!

The best you can do is use the Guest WiFi or eduroam network, if your office is close enough to one of the access points.

Warning! You can't connect directly to the obelix LSCE servers from the wifi network. You first have to connect to the ssh1.lsce.ipsl.fr gateway server, and then to obelix. More details in LSCE servers

Intranet web site: https://intranet.lsce.ipsl.fr/

There is a lot of information available on the LSCE intranet, but you can only access it from a computer connected to the LSCE network

Everything about the LSCE servers (and other software and hardware information): https://intranet.lsce.ipsl.fr/informatique/en/orme/index.php

• Reading the mails with the webmail or Thunderbird
  • As soon as you have a LSCE e-mail address, make sure it is added to you team's list
• Printers

Use the following steps:

1. Start Chrome (or try another browser)
   • it is even better if you use a Private/Incognito window to do this
2. Open the following link: https://vpn.cea.fr/
   • ignore the possible security/certificate errors
   • this link will work from the Paris Saclay network, and ANY other location (e.g. home) OUTSIDE of CEA
3. Use your initials and badge number as the login (e.g. “James Bond” ⇒ “jb007007”) and the number supplied by the activIDentity keyring, followed by your 4 digits personal code, as your password
4. Select the CEA web site you want from the “Aventail WorkPlace” portal
   • Espace Sigma: that's where you specify when you will be on holidays
   • Intranet Saclay: where you can get information about the CEA buses, CEA directory, security training registration and information, practical information, etc…
   • Intranet du voyageur: if you need to make a plane/train reservation
   • …
5. Click on logout when you are done
   • … and close your browser Incognito window if you used one

You should receive a desktop/laptop connected to the LSCE network when you arrive. Please take the time to read the instructions below, that may help make your life easier

Read the Windows 10 notes, especially the Configuring Windows 10 section

If an application requests administrator's rights to install something or make changes to the computer, and you are sure it's not a virus, use the .\admin local account and the admin password you received when your computer was configured, or ask the system administrators.

There are different window managers available, each one with different settings: you should the Cinnamon window manager (xfce is a bit too basic, and Gnome is not very convenient for doing actual work, and uses too much CPU). You can select the window manager by clicking on the little cogwheel below the password field, on the login screen.

This is Linux, you are using it for work, you should be efficient, so you should be able to do most of your work by typing commands in a terminal. The true power is in the correct use of the command-line interfaces, rather than having to spend time finding where to click in the windows. You will find some Linux documentation below.

Depending on what you need to do:

• run a single command with the root access rights
  sudo command
• become root in a terminal
  sudo su -
• use the root password when a program asks for it. Be sure the programs have legitimate reasons to ask for the root password!

Sorry, you are (almost) on your own, but a Mac is soooo easy to use, right?

Some useful mac links:

• A recommended terminal for Mac
• Using an X server on a Mac
• Mac ssh agent

If you have a Linux computer, but need to use a Windows application, follow the instructions in the Windows servers page.

If you need to use a program that is missing on the Windows cluster, see the Help section.

You have to use ssh to connect to the LSCE Linux servers, and use scp to copy files between servers

If you are really in a hurry, go directly to the Connecting to servers commonly used by LSCE users section, but it is important that you read the full ssh documentation page at least once!

Always remember that your local LSCE desktop/laptop can access remote disks on the LSCE servers.

There are lots of things you can do directly on your local computer (displaying pdf, images, using a text editor, …), rather than on the remote servers. For example, if a script running on the servers generates a pdf file, it is more efficient to open this pdf file using Acrobat/evince on your local computer, than by using evince on the server.

• The LSCE gateway: ssh1.lsce.ipsl.fr
  • Use ssh ssh1.lsce.ipsl.fr when you are outside the LSCE wired network, before accessing any other LSCE server:
    • By outside LSCE, we mean that your computer is:
      • connected to the LSCE WiFi network (instead of the wired network)
      • or physically outside LSCE
  • You can also use ssh1 when you are copying files to/from outside LSCE with scp or ssh-based tools
  • The ssh1 gateway can only be used for copying files or accessing the LSCE interactive servers!
    • You have to connect to an obelixNN interactive server before doing any real work

• The LSCE interactive servers: obelixNN (obelix2 to obelix5)
  • Use ssh obelix to access the LSCE servers from inside LSCE.
    The load balancing system will send you to the obelixNN server that has currently the smallest load.
    • By inside LSCE, we mean that:
      • your computer is connected to the LSCE wired network
      • or you are using the LSCE VPN
      • or you have a terminal connected to the ssh1 gateway
  • Never forget that you are sharing these interactive servers with other users!
    • Do not use too much CPU and/or memory for a long time. Heavy computation should be done on the LSCE cluster
    • Learn how to use the top command to determine the current load of a server and the CPU/memory usage of your processes !
  • You can use one of the following commands if you want to know on which obelixNN you are

    • $ ssh obelix
      Last login: Mon Jun  3 08:49:53 2019 from somewhere
      $ echo $HOST
      obelix5
      $ hostname
      obelix5

  • In a python script, you can use the following to get some information

    • >>> import os
      >>> os.getlogin(), os.getenv('HOST')
      ('mylogin', 'obelix5')

• The LSCE cluster (aka the batch system)
  Use this cluster for heavy duty programs, rather than killing the multi-users interactive servers
  • Heavy duty programs are programs that will prevent other users from working on the same server because these programs use a lot of CPU and/or memory for more than a few minutes. Ask your advisor, if you are not sure

• The JupyterHub notebook servers:
  • Read the Using JupyterHub servers section of the Working with Python page

• The IPSL Computing and Data Center, aka Mésocentre ESPRI (Spirit / SpiritX / HAL)
  • Note: the Spirit/X clusters have replaced the ciclad/climserv cluster (in 2022)
  • You will need a specific account (different from your LSCE account) to access these servers and use data files that are already available/mirrored there (e.g. CMIPn data).
    Do not duplicate at LSCE data already available on IPSL servers and that you can process there!
  • Requesting an IPSL account
  • IPSL ESPRI Mesocenter documentation
    • spirit clusters
  • Getting help

• The asterixNN servers
  Some wise LSCE elders may mention these servers, but they don't exist any more!

It is important to know which shell you are using if you need to configure and tune your Linux account. The shell is basically the program that waits for you to type commands in a terminal and passes them to the computer

If you are not sure which shell is running in your terminals, you can use echo $SHELL to find out:

# Somewhere with tcsh
 > echo $SHELL
 /bin/tcsh

# Somewhere else with bash
 $ echo $SHELL
 /bin/bash

If the Linux server you are working on seems slow, or if you want to get an idea of the resources you are using, you should use the top command

  >hostname
obelix3

 >top
top - 17:48:51 up 8 days, 23:43, 29 users,  load average: 1.22, 1.38, 1.98
Tasks: 324 total,   2 running, 320 sleeping,   2 stopped,   0 zombie
Cpu(s): 12.6%us,  0.0%sy,  0.0%ni, 87.4%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:  32877756k total, 14706188k used, 18171568k free,    21372k buffers
Swap: 33554428k total,  1628972k used, 31925456k free, 14105940k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
10193 john      20   0  302m 4336 2272 R 100.0  0.0   7033:08 emacs
16504 john      20   0  110m 2008 1228 S  0.7  0.0   0:00.13 tcsh
26344 jane      20   0 1635m  20m 3880 S  0.3  0.1   0:44.90 ipython
    1 root      20   0 21448 1108  892 S  0.0  0.0   0:01.61 init
    [...]

The top lines give you some summary information about the system, but you should monitor the memory usage (VIRT and RES), CPU and TIME+ columns. By default, top will will put the processes using the more CPU at the top (as shown above). You can see above that the emacs text editor has clearly crashed, because it should not use 100% CPU for such a long time

• The TIME+ information is in minutes (e.g. 5432:01 means 5432 minutes and 1 second, 25:15.20 means 25 minutes, 15 seconds and 20% of 1 second)

• Memory usage:
  • %MEM (Kb): percentage of the total available physical memory used by a process. This is based on RES, the non-swapped physical memory used by a process
  • VIRT (Kb): total amount of virtual memory (i.e. memory temporarily swapped/saved to disk) used by a process. It includes all code, data and shared libraries plus pages that have been swapped out and pages that have been mapped but not used
  • The free command can display the memory usage of a server at a given time. available is an estimation of how much memory is available for starting new applications, without swapping

    •  % free -ht
                    total        used        free      shared  buff/cache   available
      Mem:            62G        1.5G         56G         11M        5.0G         60G
      Swap:           62G        1.7G         60G
      Total:         125G        3.3G        116G

If you type M, the processes will be sorted my memory usage, as shown below

      PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
16092 bob       20   0 8488m 8.1g 7912 R 99.6 25.9   1:41.38 idl
23256 john      20   0  232m  70m 2632 S  0.0  0.2   0:01.36 R
26344 jane      20   0 1635m  20m 3880 S  0.0  0.1   0:44.87 ipython
10619 bob       20   0  954m 9292 2664 S  0.0  0.0   0:40.67 gedit
 1362 nslcd     20   0  442m 6056 2208 S  0.0  0.0   2:13.98 nslcd
10193 john      20   0  302m 4336 2272 R 99.9  0.0   7024:13 emacs
    [...]

Other useful keys:

kill proc_number

kill -9 -1

qdel

It is possible to directly access some disks on the LSCE servers from:

• a Linux or Windows computer connected directly to the LSCE ethernet/wired network
• or a remote computer using the LSCE VPN

For security reasons, only backed up disks and scratch01 can be accessed from Windows with dfs

e.g. if you have a /home/scratch01/your_lsce_login path on a Linux server, you can access it the following way from your desktop/laptop:

• Linux computer: cd /home/scratch01/your_lsce_login
• Windows computer explorer: \\dfshost\dfs\scratch01\your_lsce_login
  

• Location: the location depends on the machine you are connected to in the current terminal
  • On the LSCE servers (obelixN): /home/users/your_lsce_login
    • Non LSCE servers:
      • TGCC: /ccc/cont003/home/dsm/your_tgcc_login
        You can find out the location of all your directories with ccc_home -a (use ccc_home -h to find out all the available options)
      • ciclad: /home/your_ciclad_login
  • On a local Linux desktop (lsceNNNN): /homel/your_lsce_login
  • The configuration files of your Linux account are stored in the home directory (use ls -a to see hidden files and directories). Ask your advisor if you should copy somebody else's configuration files
    • tcsh shell users: .cshrc and .login
    • bash shell users: .profile and .bashrc
    • emacs text editor: .emacs
  • You can go to your home directory by typing cd (no arguments),
    and to the home directory of somebody else with cd ~login_name
    • At LSCE, you can't read the content of somebody else's home directory by default

• Backup of the home directory on the LSCE servers: YES! Every night
  • More details on the LSCE backup system page
  • Put the most important sources/scripts/data/papers in this directory
  • Do not use the home directory to store:
    • temporary files (use scratch01)
    • big (or many) data files (use scratch01 or a project specific disk)
  • Your home directory will be archived when your account is closed, so that you or your advisor can access it later, if required

• Backup of the TGCC home directory: YES!
  • Use the following command to access the backup: cd ~/.snapshot

• Quota: YES!
  There is a limit to what you can store in your home directory. If you reach this limit, you may experience some seemingly random errors (can't connect to the servers, the new files have a 0 size, …)
  The example below shows how to determine the quota on the LSCE disks: you can never go above the size displayed in the limit column, and you cannot stay more than 7 days above the limit displayed in the quota column! The grace column will display either None (you cannot write on the disk anymore, unless you get below the quota limit) or N days (you can use the disk, but you have N days to go below the quota limit)
  

   > quota -s
  
  Disk quotas for user johndoe (uid 6369):
       Filesystem  blocks   quota   limit   grace   files   quota   limit   grace
  prolix3:/users/   4658M   4883M   4981M           23503       0       0

• Location: /home/scratch01/your_lsce_login

• Backup: NO
  • Use this directory for temporary and big data files
  • The files older than 6 months (and possibly older than 1 month) will be automatically deleted if there is not enough free space. See the Scratch area section for more details on the retention policy
    • In other words, the files are guaranteed to stay 1 month and possibly longer

• Quota: NO

• Location: /home/some_project_name/your_lsce_login
  • Ask your advisor if you should use a specific project directory and the technical details about it
  • Use this directory for storing the (big) data files that you can't store safely in the scratch directory, and should not store in the home directory

• Backup: ask your advisor

• Quota: ask your advisor

Your Linux environment may sometimes stop working correctly because you have exceeded the allowed quota on your disks (check your quotas with quota -s). You can use the following command to get the size of all the directories and files in the current directory, sorted by size: du -sh * | sort -rh

 >cd /home/scratch01/johndoe
/home/scratch01/johndoe

 >du -sh * | sort -rh
58M     GLAC1DHiceF26.nc
47M     dashboard
44M     GLAC1DdrainagePointerF26ka.nc
15M     octcdf
12M     1901.nc
3.2M    cmake
...

Add something about df

Your advisor will let you know which (version of which) programs you should use. Some programs are available by default, and you have to use the module command to access other programs.

At any time, you can use the which command to determine where a program is located. The option to get the version information of a program may vary (launch the program without arguments, with -v or --v, …).

 > which ncdump
/usr/bin/ncdump

 > ncdump
ncdump [-c|-h] [-v ...] [[-b|-f] [c|f]] [-l len] [-n name] [-p n[,n]] [-k] [-x] [-s] [-t|-i] [-g ...] [-w] file
[...]
netcdf library version 4.3.3.1 of Dec 10 2015 16:44:18 $

 > which python
/usr/bin/python

 > python --version
Python 2.7.5

Most useful module options:

• module list: list the currently loaded modules
• module avail [mod_name]: list the available modules (or only the ones with a name starting with mod_name)
• module load module_a … module_n: load modules (automatically including dependencies)
  • module load module_a: load default version of module_a
  • module load module_a/vvv: load vvv version of module_a
• module purge: remove all modules

Detailed example:

 > which ncview
ncview: Command not found.

 > module list
No Modulefiles Currently Loaded.

 > module avail ncview
ncview/2.1.7

 > module load ncview netcdf/4

 > module list
Currently Loaded Modulefiles:
  1) ncview/2.1.7   2) netcdf/4
  
 > which ncview
/usr/local/install/ncview-2.1.7/bin/ncview

 > which ncdump
/usr/local/install/netcdf-4.3.2/bin/ncdump

• UNIX_doc_JYP.pdf: some basic Linux documentation. Make sure you read it!
  1. A list of the most useful commands
  2. A tutorial
• Useful Linux commands for working with (potentially big) text files (houchesa4_updated.pdf)
• Using ssh and scp

There are lots of text editors available on Linux computers/servers (emacs, nano, vi, vim, gvim, gedit, …). You can use them:

• In basic text terminals
  • If you only need to read a text file, use a pager like less (or more): less my_text_file.txt
    • SPACE goes down one page
    • b (back) goes back one page
    • /STRING looks for STRING in the text (then next, previous, quit)
    • q exits
  • nano is probably the easiest editor in text mode to use!
  • Useful if you are in a hurry, and/or working on a remote server with a bad network connection or have no local X server running
  • You should know the basic commands and shortcuts!

• With a full Graphical User Interface (aka GUI) allowing you to use the mouse+menus/buttons/etc…
  • you will need to have a local X server running, if you want to use the text editor GUI on a remote server

You can easily view/edit a file by just typing nano my_text_file.txt in a terminal, and looking at the instructions displayed at the bottom of the screen (e.g. ^X Exit means that you can exit the editor by typing CTRL-X)

• vi cheat sheet
• vi reference card

Notepad++ is a nice and powerful text editor for Windows, but it is also very easy to install emacs on Windows

• What are binary (binaire_jyp.pdf) and NetCDF (netcdf_jyp.pdf) files?
  • A reminder about numerical precision
  • Network Common Data Form (NetCDF) official page
    • The different netCDF formats (and do not forget to use ncdump -k data.nc if you need to know the format)
• The NetCDF file format is self-documented, and the metadata of climate data files often follows the CF (Climate and Forecast) Metadata Conventions
  • The files in the CMIPn (n ≥ 5) database are written with the CMOR 3 library in order to make sure that they follow exactly the requested standards

• Displaying the content of a NetCDF file in a text format
  Check the ncdump options. People usually want to quickly check the metadata, and use other programs to work with the actual data
  • ncdump -h some_climate_data_file.nc: only display the metadata (h_eader of the file)
  • ncdump some_climate_data_file.nc | less: dump the full content to text, and use less to display it page by page
    • Maybe it is time to learn how to work efficiently with big text files
• You can use Panoply and ncview for quickly visualizing NetCDF files
• Manipulating NetCDF files:
  • Climate Data Operators (cdo)
    • Use module load cdo on the LSCE and IPSL servers
  • netCDF Operator (NCO operators)
    • Use module load cdo on the LSCE and IPSL servers

• NetCDF CF (Climate and Forecast) Metadata Conventions

• Gridspec: A standard for the description of grids used in Earth System models
  • A nice document for discovering the different type of (ir)regular grids used in climate models!

• UGRID Conventions for storing unstructured (or flexible mesh) model data

• SGRID Conventions for storing staggered data on structured grids that are consistent with the UGRID conventions

Check the dedicated section of the Python page

• JYP's Python notes

Send a mail to help-lsce

When you need help from the administrators or other people, it will save a lot of time if you describe your problem as accurately as possible (do not just report XXXXX is not working as expected !)

You can copy-paste the error message(s). Finding the error log file(s) of the program/system (when available), or using the verbose mode of a program will make the problem resolution much easier and faster. The example below shows how to use the verbose mode of ssh

 >ssh -v obelix
OpenSSH_7.4p1, OpenSSL 1.0.2k-fips  26 Jan 2017
debug1: Reading configuration data /home/users/johndoe/.ssh/config
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: /etc/ssh/ssh_config line 58: Applying options for *
debug1: Connecting to obelix [157.136.72.75] port 22.
debug1: Connection established.
debug1: key_load_public: No such file or directory
debug1: identity file /home/users/johndoe/.ssh/id_rsa type -1
[...]
debug1: Entering interactive session.
debug1: pledge: exec
debug1: client_input_global_request: rtype hostkeys-00@openssh.com want_reply 0
debug1: Requesting X11 forwarding with authentication spoofing.
debug1: Sending environment.
debug1: Sending env LANG = en_US
Last login: Thu Mar  1 15:52:44 2018 from lsce4078.lsce.ipsl.fr

 >

Sometimes, it may also help to send a picture of what is wrong!

On Linux, you should find a screenshot application in the tools available on the system. See the example below for people using the xfce windows manager

On Windows, you can use Snip & Sketch, or the Capture screen option of XnView

Check the ESPRI User Support page

Ask you advisor, who has probably already told you everything that was listed above, so you did not really learn anything new by reading this page

[ PMIP3 Wiki Home ] - [ Help! ] - [ Wiki syntax ]