This repo is about script and resources to maintain the room-D324 computer called 'the MONSTER'

This machine runs UBUNTU (swithced from centOS on the 19th september 2022)

Every user has their own user and has root priviledge

It has two main Volumes:

• ssd for daily use (1 TB + 1 TB)

  ⚠️ for the moment, please do not use DATA_SSD

• hd for storage (6TB + 2 TB (should be) RAID 5)

• labMONSTER
  • Table of content
  • Best practices
  • New user
  • FAQ
    • How can I connect remotely?
    • I can't connect remotely
    • How do I know if there is HD space on the computer?
    • RAM get full and my command is killed
    • How do I know if someone else is using part or all the cpus/RAM?
    • I run a command in remote and closed the connection and it seems that the command was quit
    • How do I exchange files from somewhere to the MONSTER and back (especially if working remotely)?
    • Machine is stuck when switching user with blinking cursor
  • Specs
  • List installed software
    • Installed with sudo snap install ***
    • Installed with sudo apt-get ***
    • Installed manually
      • FSL
      • MATLAB 2017a
      • MATLAB 2018a
      • SPM12
      • AFNI
      • freesurfer
      • itksanp
    • to do

• Free up space asap
• Can you run the analyses you intend to perform on the UCLouvain cluster (e.g. fmriPrep, mriqc, anything that can be run on Octave)? If yes, then try to move it there.
• Delete your user when you know you will not use the monster anymore
• Book screen and/or cpu time on the calendar cpp : monster
• The monster is a common space so contribute to make it better by sharing issues/solution via this documentation
• If you are running massive computation speak with other user
• For your cyber sicurety, better not leaving logged sensitive websites eg emails, google/safari (all your passwords are saved there!!!) etc.
• Consider to share your user password with someone else in the lab so easier to deal woth your use if you are not aroudn a and you forgot to back up or delete your user
• After a year of unused user, it will be deleted automatically
• If you have problems, you need to install softwares or anything else you are not sure about, open an issue here or ask the current monster maintaner (at the moment is @marcobarilari)

1. Read these docs and ask question/open issue if something is not clear
2. Ask a sudo user to create your user
3. Enjoy and improve others experience by imporoving this documentation with your experience

Ask Marco

Are you in the UCLouvain network or using a VPN to connect as-if you are within the UCLouvain network? Check how to use UCLouvain VPN here

Check whether IP address has changed via ip route on the monster. It should be the 3rd address printed.

run df -h on a terminal, volumes of interests are:

• tmpfs: HDD for archive usage
• /dev/nvme0n1p2: SSD on which the OS is istalled and operting for day by day use.
• /dev/nvme1n1p1: avoid using this one, it will become a back-up fo /dev/nvme1n1p2 (RAID1)

Filesystem      Size  Used Avail Use% Mounted on
tmpfs           6,3G  4,3M  6,3G   1% /run
/dev/nvme0n1p2  938G  230G  661G  26% /
tmpfs            32G  150M   32G   1% /dev/shm
tmpfs           5,0M  4,0K  5,0M   1% /run/lock
/dev/nvme1n1p1  938G   28K  891G   1% /mnt/91061732-7f0e-4c15-9251-185421ec948a
/dev/nvme0n1p1  511M   31M  481M   6% /boot/efi
tmpfs           6,3G  164K  6,3G   1% /run/user/1000
tmpfs           6,3G  2,5M  6,3G   1% /run/user/1003
tmpfs           6,3G  4,7M  6,3G   1% /run/user/1001
tmpfs           6,3G  120K  6,3G   1% /run/user/1004
tmpfs           6,3G  2,5M  6,3G   1% /run/user/1005
tmpfs           6,3G   88K  6,3G   1% /run/user/127

You may need to increase the swap memory.

To increase the swap memory see here https://askubuntu.com/questions/178712/how-to-increase-swap-space

At the moemnt of this update is 200GB since we had problems with running layerification with upsampled data. The image is at /media/fasthdd/swapfile.img, to delete and reset the swap in a different way:

# swithc off the swap space
sudo swapon /media/fasthdd/swapfile.img

# delete the image
sudo rm /media/fasthdd/swapfile.img

# follow back the instructions at the link above

NB: the swap memory is reset to 2GB at each reboot. It needs to reset to the actual swapfile.img back to serve as swap.

run htop on the terminal, it is a "graphic" interface displaying live command running and by who, each cpu usage, and RAM usage.

Did you use screen? Check the guide SSH-tips

Please use Datalad and GIN. Why? Because it is easy for you to drop files and free up space since this machine is a crunch machine and not a data archive machine. BUT... If you are in a rush, check SSH-tips on how to use scp command.

use Ctrl + Alt + F2

(not really working properly....)

ADD SPECS

to update all the software use sudo snap refresh

• gitkraken - available for all users
• vscode - available for all users
• octave (7.1.9) - available for all users
• docker (20.10.14) - available for all users, docker images are shared across users

[ TIPS 💡 ]

If ayou are a new user and don't want to use 'sudo' everytime you call a docker command , use the first time sudo usermod -aG docker $USER (log out and in again :/)

• git - available for all users
• git-annex - available for all users
• openssh-server (to acces remotly) installation guide and user guide- TO CHECK if available for all users
• screen (to open terminal sessions while accessing remotley) installation guide and user guide- TO CHECK if available for all users
• r - TO CHECK if available for all users

[ TIPS 💡 ]

1. to update the list of available softwre online sudo apt-get update
2. to upgrade the list of software installed via apt use sudo apt-get upgrade

• chrome - TO CHECK if available for all users
• dropbox - TO CHECK if available for all users
• conda via miniconda - each user has to install it indipendently (see here miniconda installer)
• rstudio - TO CHECK if available for all users
• datalad - each user has to install it indipendently via pip install

fsleyes via conda - each user has to install it indipendently

TO CHECK if available for all users

[ TIPS 💡 ]

• add this in in ~/.baschrc to help FSL being more reachable

#FSL
FSLDIR=/usr/local/fsl
. ${FSLDIR}/etc/fslconf/fsl.sh
PATH=${FSLDIR}/bin:${PATH}
export FSLDIR PATH

• open FSL gui using fsl in the terminal

• open fsleyes gui using fsleyes in the terminal, if installed via conda you need to activate it first via conda activate

[ TIPS 💡 ]

• open matlab gui using /usr/local/MATLAB/R2017a/bin/matlab in the terminal, to be faster add an aliasia in the ~/.bashrc file

alias matlab2017=/usr/local/MATLAB/R2017a/bin/matlab

[ TIPS 💡 ]

• open matlab gui using /usr/local/MATLAB/R2018a/bin/matlab in the terminal, to be faster add an aliasia in the ~/.bashrc file

alias matlab2018=/usr/local/MATLAB/R2018a/bin/matlab

• how to run matlab script the lazy way:

1. Write down a bash function in eg ~/.my-func named matlabrun.sh

#!/bin/bash

#matlabrun function 

matlabrun () {

    script_path=$1

    matlab -nodisplay -nosplash -nodesktop -r "run('"${script_path}"');"

}

where matlab here is an alias for /usr/local/MATLAB/R2018a/bin/matlab

2. Run or add source ~/.my-func/matlabrun.sh to either ~/.bashrc ~/.bash_profile or ~/.zrsrc depending on your set up environment

3. Usage in a terminal:

matlabrun path/to/script.m

each user has to install it indipendently

[ TIPS 💡 ]

The gui might not work properly, solution: add this snippet to you .zshrc / .bashrc / etc.

# make sure spm gui works
export SPM_HTML_BROWSER=0

each user has to install it indipendently (most of missing libraries should be installed for everyone so now should be easier)

[ TIPS 💡 ]

• afni setup is not 100% done since we are missing some libraries as libGLw.so.1 and libgsl.so.19

from running 'afni_system_check.py -check_all':

testing ability to start various programs...
    afni                 : success
    suma                 : FAILURE
        suma: error while loading shared libraries: libGLw.so.1: cannot open shared object file: No such file or directory
    3dSkullStrip         : FAILURE
        3dSkullStrip: error while loading shared libraries: libGLw.so.1: cannot open shared object file: No such file or directory
    uber_subject.py      : success
    3dAllineate          : success
    3dRSFC               : FAILURE
        3dRSFC: error while loading shared libraries: libgsl.so.19: cannot open shared object file: No such file or directory
    SurfMesh             : FAILURE
        SurfMesh: error while loading shared libraries: libGLw.so.1: cannot open shared object file: No such file or directory
    3dClustSim           : success
    3dMVM                : success

available for all users

[ TIPS 💡 ]

• set up freesurfer by adding this in ~/.bashrc and then take care to have license somewhere

export FREESURFER_HOME=/usr/local/freesurfer/7.3.2
source $FREESURFER_HOME/SetUpFreeSurfer.sh

available for all users

• run this in the terminal to call the app via terminal with itksnap

export PATH=$PATH:/usr/local/itksnap-3.8.0-20190612-Linux-gcc64/bin