Set up multi-factor authentication using a token app or SMS.
Visit www.tap.tacc.utexas.edu and use the following information:
- System: Lonestar6
- Application: DCV remote desktop
- Project: IBN22012
- Queue: vm-small, development, normal, gpu, gpu-development
- Time limit: 48:0:0 (Small and normal queues have time limits of 48h, development queue has a time limit of 2h)
- Click submit and connect once you’re off the queue
- Use Utilities to link scratch work here
- Type:
ssh [email protected]
- Confirm with:
yes
- Enter your TACC password
- Enter your duo mobile key code
- Get idev node before doing anything here
- Type:
idev -p NODETYPE -t TIME(h:m:s)
- Put all working files in
/scratch
, there's no space limit- BUT it purges files if you don't access them for 10 days. Always back up files to
/work
or/corral
.
- BUT it purges files if you don't access them for 10 days. Always back up files to
/work
has a limit of 1024 GB. If work is full then transfer to/corral
. See corral_guide.- Type
cd
for$HOME
,cdw
for$WORK
,cds
for$SCRATCH
.- In shell, type
$HOME/
and press tab, it will show you the absolute path. You need the absolute path instead of$SCRATCH
to access directories when you transfer files from and to local computer and when you use jupyter notebook.
- In shell, type
rsync -avz SOURCE DESTINATION
SOURCE
is where you’re sending the file fromDESTINATION
is where you’re sending the file to
- No sudo, yum, apt-get or anything similar.
- The bash file is located in the
home1
folder. Usels -a
to see it in the directory,nano
to edit it.- There should be this command in the last line of the file:
conda deactivate
- There should be this command in the last line of the file:
- Include the Anaconda installation in the bash.
- Add directories to the PATH:
export PATH=.:$WORK/YOUR/ANACONDA/PATH:$PATH
. Do this every time you install/compile a new thing. - Source the bashrc when you added a new path:
source ./bashrc
.
- Download through the web browser, making sure it’s installed under
$WORK
not$HOME
.
- Install libmamba in the base environment and set it as solver. More information can be found here.
- Add conda-forge in the environment channels. More information can be found here.
- Other environments: Sklearn (optional)
- Add minian and other environments as jupyter kernel:
python -m ipykernel install --user --name=EnvName
. More information can be found here.
- Update GCC in minian environment:
conda install -c conda-forge gcc=12.1.0
- Downgrade pyviz:
conda install pyviz_comms==2.2.0
- Use new version of pipeline from minian github (
minian-install --notebooks
), but remember to change the custom codes for slicing and parameter defining. - Add to PATH
"/work/09117/xz6783/ls6/Anaconda/envs/minian/bin"
. - Need to install ffmpeg, which requires nasm and other things
- You might need to get a new load after these steps since jupyter only use the
$PATH
when it started and can’t be updated during running.
From browser: Change firefox download file to work or scratch first!!
From wget
or Curl
(need to compile): cd to target folder first
Unzip:
- First download the Unzip package to work. Then compile it in the terminal (at the downloaded folder) following the instruction here
- After installation, unzip the file in the terminal by running “unzip filename”
- You can also use 7zip which might be better
https://wikis.utexas.edu/display/CoreNGSTools/Linux+fundamentals
https://issm.ess.uci.edu/trac/issm/wiki/lonestar
https://filezilla-project.org/download.php?type=client
https://www.linux.com/training-tutorials/get-know-rsync/
https://wikis.utexas.edu/display/CoreNGSTools/Day+1%3A+Intro+to+NGS%2C+Linux+and+TACC
https://tacc.github.io/ctls2017/docs/intro_to_linux/intro_to_linux_05.html
https://tacc.github.io/ctls2017/docs/intro_to_linux/intro_to_linux_01.html
https://sites.google.com/site/tacchadoop/home/upload-your-data-to-tacc-computer