Compute Cloud

Notice

In case you encounter any (severe) problem with the LRZ Compute Cloud please refer to the  LRZ status page first to get more information.

The LRZ Compute Cloud

The LRZ Cloud portal is a service offering from LRZ for its customers. The idea is to provide a responsive service that allows customers to upload and use their own virtual machine images.  The LRZ Compute Cloud system can be accessed via the login page

https://cc.lrz.de

Target Audience

The target audience of the Compute Cloud are users that need to get resources for run services and test software quickly. Users can create VMs on their own with their preferred operating system (except Windows) and being root in their VMs: VMs can be configured as needed, users are free in installing the software they need but they are also responsible for everything happening in their VMs.

The Compute Cloud can be run to perform calculations, host testing environments, and can accompany lectures and workshops. The cloud can also be used to run services for research projects. It is operated following the best effort principle.
Our resources are limited.

There are some important points you should be aware of:

  • Data Privacy Policy / Datenschutz
    Die Nutzung der LRZ-Infrastruktur zur Verarbeitung von medizinischen Forschungsdaten, lt. Art. 9 DSGVO „Verarbeitung besonderer Kategorien personenbezogener Daten“, ist grundsätzlich ausgeschlossen. Allerdings sind Ausnahmen möglich bei Abschluss einer Sondervereinbarung zur existierenden Auftragsverarbeitungsvereinbarung (AVV) mit der LMU/TUM, und andererseits ist in der Regel die Installation und der Betrieb einer eigens abgesicherten Infrastruktur erforderlich. In diesem Fall wenden Sie sich bitte mit Ihrem Anliegen an die LRZ-Leitung: leitung@lrz.de.
  • We do not offer support for problems with software inside your VMs!
    The Compute Cloud is an OpenStack-based Infrastructure-as-a-Service system. We try to make to sure the infrastructure works properly, but what you do on this infrastructure is your responsibility. We are happy to help with any issues directly related to the system, but we do not have the capacity to impart Linux admin or OpenStack essentials via the ticket system or email. This kind of knowledge is a requirement for the successful usage of this service. If you do not possess it, then maybe you find a suitable LRZ course, and otherwise there are tons of tutorials about all these topics online.
  • The Compute Cloud is not an HPC system!
    In contrast to other server hosting systems like our VMware we do also offer large flavors which can be used to run simulations and computations, but this is still a cloud and not a cluster. The virtualization, random vCPU-pCPU-allotment and CPU overcommitment can have a noticeable negative impact on especially CPU-, cache-, or memory-intensive workloads. If you want maximum performance, then one of the other HPC systems is a much more suitable option.
  • We do not offer a storage system for large datasets!
    Please note that the storage backend of the Compute Cloud is used to host the virtual disks belonging to the VMs in the cloud. It is not meant to store large data sets. If you have huge amounts of data you want to process within VMs in the Compute Cloud please use LRZ's Data Science Storage service.
  • We do not create backups!
    You are responsible for backing up your data. We cannot create volumes that have been (accidentially) deleted or restore corrupted/deleted files within a single volume. If you need backups you can backup the relevant data with rsync or the tool(s) of your choice or use LRZ's Backup service.
  • We do not offer high availability!
    The Compute Cloud is not suitable to host services that need to be highly available although we do our best. If you need high availability, please do not use the Compute Cloud but consider using our Server-Hosting offering.
  • We do not offer any guarantees that you can access the resources you want!
    As already stated above: Our resources are limited. If all resources are fully used by other users there is no way to create new VMs. We try to mitigate this by assigned default quotas to every user that prevents a user do consume too many resources at once. However, since the number of users is increasing we cannot guarantee that you always get what you want. One solution to that is that you buy your own Cloud hardware (called "Attended Cloud Housing"):
    If you have money to spend and want to have exclusive access to Cloud resources please drop us a line.
  • Make sure to follow our security considerations!
    When using a publicly reachable server you should have in mind that you need to invest some effort to secure your machine to prevent it from being highjacked or hacked. We provide some ideas on how to secure some of the basic pain points in our Security Considerations document.
    Remember: You are responsible for your VMs.
  • We do not support IPv6!
    The IPv6 support of the OpenStack version we are using is not sufficient for us to consider it a stable feature. Therefore it is disabled in the Compute Cloud.

Billing and Budgeting

Using the public part of the LRZ Compute Cloud is subject to a fee.

Refer to the budgeting section for further information.

Getting access

To access our cloud environment, you need a LRZ project that is enabled for the LRZ Compute Cloud service. More information on this subject can be found here.

Attended Cloud Housing

We offer the possibility to buy hardware on a behalf of a user and provide exclusive access to these resources using the usual LRZ Compute Cloud interface. You can find more information on this offering here.

Frequently Asked Questions (FAQ)

The FAQ section covers questions that will come up regularly. You should consult this page before contacting the cloud support team.

When you contact us please always provide the following information, this will help us a lot to speed up the process:

  • your User Name (LRZ Kennung) you use to login to the cloud
  • the UUIDs of the cloud components, i.e. VM, instance, volume, network etc. which cause the problems.

Tutorials

Announcements

Hardware Overview

Node TypeNumber of Nodes

CPU & GPU Cores /

per Node

RAM (GB) /

per node

Access
Compute Node8240x Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz384 GB

Up to 10 CPUs: all cloud users.

More than 10 CPUs: restricted, contact us: cloud support team

GPU Node32

40x Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz 

2x Nvidia Tesla V100 16 GB RAM

768 GB

restricted, contact us: cloud support team

Huge Node1192x Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz6.000 GB

restricted, contact us: cloud support team

Details about the available Flavors can be found here: Flavors

Policies