Project

General

Profile

New cluster » History » Version 10

Martin Kuemmel, 05/12/2022 09:39 AM

1 2 Martin Kuemmel
h1. New computing cluster in Koenigstrasse
2 1 Martin Kuemmel
3 1 Martin Kuemmel
h2. Introduction
4 1 Martin Kuemmel
5 1 Martin Kuemmel
Since January 2022 we have a new computing cluster which is installed int he server room of the physiscs department at Koenigstrasse. Temporarily attached to the cluster is a 10TB disk for processing. We are currently (17th March 2022) waiting for a large amount of storage (40TB) which will then replace this temporary solution.
6 1 Martin Kuemmel
7 3 Martin Kuemmel
h2. Hardware
8 3 Martin Kuemmel
9 3 Martin Kuemmel
* there are in total 8 compute nodes avalable;
10 3 Martin Kuemmel
* the compute nodes are named "usm-cl-bt01n[1-4]" and "usm-cl-bt02n[1-4]";
11 3 Martin Kuemmel
* each node has 128 cores;
12 3 Martin Kuemmel
* each node has 500Gb available;
13 3 Martin Kuemmel
14 1 Martin Kuemmel
h2. Login
15 1 Martin Kuemmel
16 1 Martin Kuemmel
* public login server: login.physik.uni-muenchen.de;
17 1 Martin Kuemmel
* Jupyterhub: https://workshop.physik.uni-muenchen.de;
18 1 Martin Kuemmel
* both the server and the Jupyterhub require a two-factor-authentication with your physics account pwd as the first authentication. Then you can use a smartphone app like Google Authenticator (or any other app that generates time-based one-time-passwords). The app needs to be registered here: https://otp.physik.uni-muenchen.de, it is there called a soft-token.
19 1 Martin Kuemmel
20 1 Martin Kuemmel
Graphical access to the cluster or its login nodes is possible, and I am currently trying to figure out the most efficient way for me.
21 1 Martin Kuemmel
22 1 Martin Kuemmel
h2. Processing
23 1 Martin Kuemmel
24 1 Martin Kuemmel
* as on our local cluster "slurm" is being used as the job scheduling system. Access to the computing nodes and running jobs requires starting a corresponding slurm job;
25 1 Martin Kuemmel
* the partition of our cluster is "usm-cl";
26 1 Martin Kuemmel
* from the login node you can start an interactive job via "intjob --partition=usm-cl" (additional slurm arguments are accepted as well);
27 8 Martin Kuemmel
* I created a "python script":https://cosmofs3.kosmo.physik.uni-muenchen.de/attachments/download/285/scontrol.py which provides information on our partition (which jobs are running on which node, the owner of the job and so on);
28 5 Martin Kuemmel
* I have also put together a rather silly "slurm script":https://cosmofs3.kosmo.physik.uni-muenchen.de/attachments/download/283/test.slurm
29 1 Martin Kuemmel
30 1 Martin Kuemmel
h2. Disk space
31 1 Martin Kuemmel
32 1 Martin Kuemmel
* users can create their own disk space under "/project/ls-mohr/users/" such as "/project/ls-mohr/users/martin.kuemmel";
33 1 Martin Kuemmel
34 1 Martin Kuemmel
h2. Installed software
35 1 Martin Kuemmel
36 1 Martin Kuemmel
We use a package manager called spack to download and install software that is not directly available from the linux distribution. To see what is already installed, do the following on a computing node:
37 1 Martin Kuemmel
38 1 Martin Kuemmel
* "module load spack"
39 1 Martin Kuemmel
* "module avail"
40 1 Martin Kuemmel
41 1 Martin Kuemmel
Adding more software is not a problem.
42 9 Martin Kuemmel
43 10 Martin Kuemmel
h2. Euclid processing on the cluster
44 10 Martin Kuemmel
45 10 Martin Kuemmel
While OS, libraries and setup is different from EDEN-?.?, it is possible to load and run in an EDEN-3.0 environment using a container solution. The cluster offers "singularity":https://sylabs.io/guides/3.0/user-guide/quick_start.html as a container solution. While singularity is not officially supported in Euclid, it is being used in a limited role, and singularity is able to run docker images, which is the supported container format in Euclid. To work in an EDEN-3.0 on the new cluster you need to get the docker image doing:
46 10 Martin Kuemmel
* load singularity via:
47 10 Martin Kuemmel
  <pre>
48 10 Martin Kuemmel
  $ module load spack
49 10 Martin Kuemmel
  $ module load singularity</pre> Note that the singularity version which is directly available on the computing nodes at "/usr/bin/singularity" does *not* work. The correct version loaded via the modules is at "/software/opt/focal/x86_64/singularity/v3.8.1/bin/singularity".
50 10 Martin Kuemmel
* pull the Euclid docker image via: <pre>singularity pull --docker-login docker://gitlab.euclid-sgs.uk:4567/st-tools/ct_xodeen_builder/dockeen</pre> With the gitlab credentials the docker image is stored in the file "dockeen_latest.sif"
51 10 Martin Kuemmel
52 10 Martin Kuemmel
The docker image can be run interactively:
53 10 Martin Kuemmel
 <pre>$ singularity run --bind /cvmfs/euclid.in2p3.fr:/cvmfs/euclid.in2p3.fr --bind /cvmfs/euclid-dev.in2p3.fr:/cvmfs/euclid-dev.in2p3.fr <path to/>dockeen_latest.sif</pre>
54 10 Martin Kuemmel
It is also possible to directly issue a command in EDEN-3.0:
55 10 Martin Kuemmel
 <pre>$ singularity run --bind /cvmfs/euclid.in2p3.fr:/cvmfs/euclid.in2p3.fr --bind /cvmfs/euclid-dev.in2p3.fr:/cvmfs/euclid-dev.in2p3.fr <path to/>dockeen_latest.sif  <command_name></pre>
56 10 Martin Kuemmel
In both cases the relevant EDEN environment must first be loaded with:
57 10 Martin Kuemmel
<pre>
58 10 Martin Kuemmel
$ source /cvmfs/euclid-dev.in2p3.fr/CentOS7/EDEN-3.0/bin/activate
59 10 Martin Kuemmel
</pre>
60 10 Martin Kuemmel
61 10 Martin Kuemmel
Information on the usage of singularity in Euclid is available at the "Euclid Redmine":https://euclid.roe.ac.uk/projects/codeen-users/wiki/EDEN_SINGULARITY.
62 10 Martin Kuemmel
63 9 Martin Kuemmel
h2. Support
64 9 Martin Kuemmel
65 9 Martin Kuemmel
Support is provided by the IT support (Rechnerbetriebsgruppe) of the LMU faculty of physics with the helpdesk email: helpdesk@physik.uni-muenchen.de. Please keep Joe Mohr and me (Martin Kuemmel: mkuemmel@usm.lmu.de) in the loop such that we can maintain an overview on the cluster performance.
Redmine Appliance - Powered by TurnKey Linux