HPC IT Helpdesk¶
Additional resources
Before contacting the helpdesk, please consider to get help in the HPC Talk BIH HPC user self-help forum!
Getting Help
Our helpdesk can be reached via email to hpc-helpdesk@bih-charite.de. Please read our guide on how to write good tickets first.
Please also use the handy figure below on general problem resolution. Note, it currently misses the really important HPC Talk somewhere between docs and writing to helpdesk.
Helpdesk Scope¶
Our helpdesk can support you in the following areas:
- Managing users, groups, projects and quotas on the cluster.
- User specific problems or questions regarding connections to the cluster (if potenially sensitive information involved).
- Data specicic problems or questions regarding access and permissions (if potenially sensitive information involved).
We will try our best to resolve these issues. Please note that all other questions can only be answered in a "best effort way".
Helpdesk Non-Scope¶
The following topics are out of scope for the BIH HPC Helpdesk since they are of more general interest and discussions may benefit from other user's experience as well as contribute to internal (HPC Talk) and external community knowledge:
- Generic Linux or programming questions (try stackoverflow.com).
- Generic help with Snakemake or other workflow engines (see Stackoverflow for getting help with Snakemake).
- Help with bioinformatics or other scientific software. Please contact the authors/communities of these software for help (also known as "upstream").
- Problems/questions with using the cluster scheduler or operating system (try HPC Talk).
- Problems with running your software that works in other environments (try HPC Talk).
- Requests for the installation of common software (prefer using Conda).
- You need a generic dependency that you cannot install via conda or on your own. Please read the section Administration-Provided Software to learn about the kinds of software that we will install and the kinds that we will not.
- Something that breaks from one week to another without you changing anything and you assume a change on the cluster (try HPC Talk first).