October 2021 - Users - The Open Data Hub List Archives

Questions about components deployment in detail and spark on OCP.

by Ki Dong Lee

Hi everyone, I am new to ODH. It seems that all the components in ODH will be deployed with kubeflow operator and kustomize manifests. Could you tell me how to deploy such components in detail on OCP in ODH? Another question is about spark on kubernetes. I have noticed that in ODH, If you want to deploy spark thrift server as hiver server2 , spark cluster needs to be deployed on OCP in ODH beforehand. I think, there is a way to submit spark thrift server on kubernetes/OCP <https://spark.apache.org/docs/3.0.3/running-on-kubernetes.html> directly without having spark cluster deployed on OCP. Is there any reason to do so? Cheers, - Kidong Lee.

2 years, 6 months

3
2
0 / 0

cuda notebook images build without pypi access

by Atahan Kara (Ibtech-Inf Open Solutions)

Hi everyone, I'm trying to build gpu enabled notebook images on Openshift. I don't have access to pypi directly. I am using artifactory with remote pypi repo configured. When I add -cuda overlay and include config and populate PIP_INDEX_URL environment variable, build chain is running until cuda images. Those builds like s2i-minimal-notebook-gpu uses micropipenv. Pip can be upgraded to latest and micropipenv can be installed from local artifactory. But unfortunately micropipenv tries to get packages from pypi repositories. I've tried using ENABLE_PIPENV and removed ENABLE_MICROPIPENV environments from buildConfig but we've some other issues with this config. All ideas are welcome, Ata Bu e-posta'n?n i?erdi?i bilgiler (ekleri dahil olmak ?zere) gizlidir. Onay?m?z olmaks?z?n ???nc? ki?ilere a?iklanamaz. Bu mesaj?n g?nderilmek istendi?i ki?i de?ilseniz, l?tfen mesaj? sisteminizden derhal siliniz. IBTech A.?. bu mesaj?n i?erdi?i bilgilerin do?rulu?u veya eksiksiz oldu?u konusunda bir garanti vermemektedir. Bu nedenle bilgilerin ne ?ekilde olursa olsun i?eri?inden, iletilmesinden, al?nmas?ndan, saklanmas?ndan sorumlu de?ildir. Bu mesaj?n i?eri?i yazar?na ait olup, IBTech A.?.'nin g?r??lerini i?ermeyebilir. The information contained in this e-mail (including any attachments)is confidential. It must not be disclosed to any person without our authority. If you are not the intended recipient, please delete it from your system immediately. IBTech A.S. makes no warranty as to the accuracy or completeness of any information contained in this message and hereby excludes any liability of any kind for the information contained therein or for the information transmission, reception, storage or use of such in any way whatsoever. Any opinions expressed in this message are those of the author and may not necessarily reflect the opinions of IBTech A.S.

2 years, 6 months

1
0
0 / 0

Happy to meet Open Data Hub.

by Ki Dong Lee

Hi everyone, I am happy to meet Open Data Hub, because I have also open sourced such platform, called DataRoaster(https://github.com/cloudcheflabs/dataroaster) recently. DataRoaster is a tool to provide data platforms running on kubernetes with ease and quickly. I think, the component of services provided by DataRoaster seems to be similar with ODH's. Since this October 1., I am working at Red Hat Korea, and I have been looking for the communities about data platform running on kubernetes. I hope that I will have a good chance to learn from ODH and ask questions. Cheers, - Kidong Lee.

2 years, 6 months

2
2
0 / 0

ODH + Kubeflow authentication architecture

by Ke Zhu

According to user authentication in the context of servicemesh 2.0, I'm using approaches from the article Restricting User Access via Service Mesh 2.0 with RedHat Single Sign On<https://cloud.redhat.com/blog/restricting-user-access-via-service-mesh-2....> (implementation from the repo https://github.com/raffaelespazzoli/kubeflow-ocp with some tweaks). it works great. If there's any design/architecture around the topic ODH + Kubeflow authentication for discussion or pilot?

2 years, 6 months

2
1
0 / 0

2024

2023

2022

2021

2020

Users October 2021