Migration from biocluster
If you are reading this you are a veteran from biocluster (or just a compulsive documentation reader). To see what has changed from biocluster to marbits you can read here.
Data
The first time you log into marbits you’ll notice that your /home
is empty except for a link to your personal directory in the lustre filesystem and a few hidden configuration files (see them with ll -a
).
If you cd
to the link with your user name you’ll see several directories that may sound familiar to you:
gendata
genwork
genwork/scratch
home
scratch
scratch2
scratch3
This absolute nightmare is how the storage volumes were set up in biocluster. This directory tree has been re-created within your personal directory including your old data. You should re-organize your stuff there to better suit your needs (or leave it as is if you like mental pain). Don’t be afraid of deleting things or directories there, as each user has their own copy of the directory tree. That is yours to do whatever you want. And now that you are at it, please delete all the things you don’t need anymore.
My recommendation is that you make a few meaningful directories (projects, data….) and put them right in /mnt/lustre/bio/users/your-user-name
. Then go to your /home
and build as many links as you need, pointing to the newly created folders.
Configuration files
You can use your /home
in marbits also to store your configuration files. Initially you have just a few, mainly .bashrc
with some pre-configured aliases (ll
, sq
, sqa
, si
).
.bash_history
.bash_profile
.bashrc
.emacs
.ssh
Feel free to edit them if you know what you are doing. Generally, you don’t need to add PATH items in your .bashrc
or .bash_profile
as you did in biocluster, as all software should be loadable by modules (see here).
You may add configuration files for tmux, vim, git as you need. Some example files that work well for these programs can be found here.