Login¶
Learning objectives
- Practice using the UPPMAX documentation
- Understand that there are two environments
- Understand that after login, one is on a login node
- Understand what a login node is
- Understand that on a login node, one only runs light commands
For teachers
Teaching goals are:
- Learners have practiced using the UPPMAX documentation
- Learners have read and heard that there are two environments
- Learners have read and heard that after login, one is on a login node
- Learners have read and heard about what a login node is in general terms
- Learners have practiced understanding what one runs on a login node
Lesson plan:
gantt
title Login
dateFormat X
axisFormat %s
section First hour
Prior : prior, 0, 5s
Present: present, after prior, 2s
%% It took me 14 mins, here I do that time x2
Challenge: crit, challenge, after present, 28s
%% Here I use the same time it took me to give feedback
Feedback: feedback, after challenge, 14s
Prior questions:
- What is 'HPC' in 'HPC cluster'?
- What is 'cluster' in 'HPC cluster'?
- Where are the UPPMAX HPC clusters?
- When does one use an HPC cluster?
- How does using an HPC clusters look like?
- How does one tell the UPPMAX HPC clusters what to do?
- Who has already logged in to an HPC cluster?
- Who has already logged in to Rackham?
A human taking care of the Rackham UPPMAX HPC cluster
Why?¶
To use any UPPMAX cluster, one needs to log in. In this session, we learn there are multiple ways to login. After logging in, one wants to behave considerate towards other users.
Prerequisites¶
It is assumed you have fulfilled all prerequisites for using Rackham.
Introduction¶
In this session, we warm up for using the UPPMAX HPC cluster called Rackham. Here we introduce the first terms, such as 'HPC cluster'.
We will practice using the UPPMAX documentation, so that you are confident that you can find what you need and understand what is written. Also, it helps UPPMAX staff find out how good our documentation is :-)
Exercises¶
Need a video?
Exercise 1: get an UPPMAX 2FA¶
Go to the UPPMAX documentation at https://docs.uppmax.uu.se, then answer these questions:
- Where is it is described how to setup a 2FA for UPPMAX?
- Which page does it link to?
- Setup your UPPMAX 2FA
Answer
A video can be found here
Note that it takes 15 minutes before you can use this 2FA.
Exercise 2: general understanding¶
Go to the UPPMAX documentation at https://docs.uppmax.uu.se, then answer these questions:
- Where is it is described how to log in to Rackham?
Answer
It is at https://docs.uppmax.uu.se/getting_started/login_rackham/.
You can reach this page by searching for 'Login Rackham' or 'Log in Rackham'.
- How many environments can one log in to? How do they look like?
Answer
Two:
-
a remote desktop environment, which looks like a regular Linux computer with the Xfce desktop environment installed
-
a console environment, which looks like text. No graphics. Text.
These answers can be found at https://docs.uppmax.uu.se/getting_started/login_rackham/.
- The documentation of how to log in to Rackham states: 'After logging in, one is on a login node'. What is a login node? Find it in the UPPMAX documentation
Answer
It is at https://docs.uppmax.uu.se/cluster_guides/login_node/.
You can reach this page by searching for 'Login node' or click on the link in the documentation
- What is a login node?
Answer
From https://docs.uppmax.uu.se/cluster_guides/login_node/:
A login node is the computer where you arrive after logging in to an UPPMAX HPC cluster
-
Should the examples below be done on the Rackham login node? What is your reasoning?
- Create a directory
- Copy a dataset of 100 gigabytes to another folder
- Download a dataset of 100 gigabytes
- Edit a file using an IDE such as RStudio, VSCode or VSCodium
- Edit a file using a simple text editor such as
emacs
,nano
orvim
- Run a bash/Nextflow/Python/R script
Answers
- Create a directory
Can be done on the Rackham login node, as this is a very light calculation that takes a very short time to run.
- Copy a dataset of 100 gigabytes to another folder
Can be done on the Rackham login node, as this is a light calculation that takes a reasonable short time to run.
- Download a dataset of 100 gigabytes
Can be done on the Rackham login node, as this is a light calculation that takes a reasonable short time to run.
- Edit a file using an IDE such as RStudio, VSCode or VSCodium
Don't do this on a Rackham login node. These program need a lot of CPU and memory and are typically used for a long time.
Instead, use an interactive session.
- Edit a file using a simple text editor such as
emacs
,nano
orvim
Can be done on the Rackham login node, as this is such a light calculation that it can be run for a long time.
- Run a bash/Nextflow/Python/R script
This depends on what the script does: if it does a light calculation or takes a reasonable short time to run, you can run it on a login node.
Else, you will have to use the Slurm job scheduler.
Done?
- If you want to know what an HPC cluster is in more detail, read up on the UPPMAX clusters
- Else move on to the next session