HTCondor in Kubernetes: High Throughput Computing Overview

htcondor in k8s and k8s in htcondor n.w
1 / 26
Embed
Share

"Explore the integration of HTCondor in Kubernetes environment for high throughput computing. Learn about the benefits, challenges, and potential solutions offered by Kubernetes in this setup."

  • HTCondor
  • Kubernetes
  • High Throughput Computing
  • Integration
  • Overview

Uploaded on | 2 Views


Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.

E N D

Presentation Transcript


  1. HTCondor in K8s and K8s in HTCondor Center for High Throughput Computing

  2. The view as of Yesterday 2

  3. Why Linux 1) Our users are there 2)Good enough 3

  4. Unix Spelled unix , but pronounced? 4

  5. Unix Spelled unix , but pronounced? U - Fix 5

  6. Who is doing the Fixing? Partially HTCondor software Partially the site admin (SRE) Mainly because Linux is single-node 6

  7. Enter Kubernetes Many users are here or working to getting here K8s as distributed operating system Can k8s fix some things we had to fix before? 7

  8. Summary of Kubernetes Container "Orchestrator" Sets of containers as pods Service focused Manages network and storage 8

  9. More Kubernetes overview Central database holds all objects Pods, services, storage, nodes All objects described in yaml One command kubectl interacts with k8s

  10. Deploying Kubernetes Local or in the cloud Most deployments are cloud based could be universal cloud interface ephemeral condor pools bespoke, Local deployment is a lot of work many distros 10

  11. Cloud deployment can be easy

  12. Two projects Deploying HTcondor in k8s k8s gahp 12

  13. 1st integration Htcondor pool inside k8s E.g. Running a whole condor pool as service 13

  14. Launching daemons as pods One pod per service CM execute execute execute schedd

  15. This requires docker images So, we built some $ docker run htcondor/cm $ docker run htcondor/execute $ docker run htcondor/submit

  16. This requires docker images $docker run htcondor/cm:8.9.6-el7 And we'll be releasing You can rebase on your own distro favs.

  17. And a minicondor $ docker run t I htcondor/mini Very handy for testing and debugging

  18. Demo time

  19. 2nd integration k8s grid type E.g. Submitting jobs as pods 19

  20. K8s grid types warnings K8s is service centric, no one-to-one mapping pods -> jobs Best used for glideins or "livestock" jobs 20

  21. K8s as HTCondor grid type apiVersion: v1 kind: Pod metadata: name: job spec: . Kubernetes as grid type: Universe = grid Grid_type = k8s hostname k8s_image = my_image queue Schedd

  22. One scheduler K8s sees jobs 22

  23. K8s add submit attrs k8s_image k8s_namespace k8s_priority Idea is that the ce adds these to localize pilot Other attributes translated: RequestMemory 23

  24. K8s gahp Shipping in 8.9.6 Still early missing file tra sfer 24

  25. Demo time

  26. Thank you Questions? Please see YouTube Tutorials Center for High Throughput Computing 26

More Related Content