Welcome to the latest installment of Arrikto’s Kubeflow tips and tricks blog! In a simple Q&A format, we aim to provide tips and tricks for the intermediate to advanced Kubeflow user. Ok, let’s dive in.
What is the difference between create_component_from_func and func_to_container_op?
They are equivalent, but func_to_container_op is deprecated in v2 of the SDK. For more information on what has been deprecated see here:
https://github.com/kubeflow/pipelines/issues/6133
Credit: Alexandre Brown in Community Slack
What is in the Kubeflow 1.4.1 Patch release and why is it important?
The Kubeflow 1.4.1 Patch release was cut in response to an issue with Istio returning 404 errors. The bug could be triggered based on the order in which Knative and Istio Gateway service objects were applied.
The 1.4.1 Patch release can be found here:
https://github.com/kubeflow/manifests/releases/tag/v1.4.1
Why is my MPI Operator stuck in a CrashBackoffLoop in Kubeflow v1.2?
The MPI deployment uses the `latest` tag on the deployment, which worked when 1.2 was released but now does not. To fix this- update the tag from `latest` to `v0.2.3` either using Kustomize OR by hardcoding it in the manifest.
Credit: Jesse Wolf in Community Slack
Is there a work around for cloud-endpoints-controller being stuck in ImagePullBackoff?
If you go into the logs you will see something along the lines of:
> e "gcr.io/cloud-solutions-group/cloud-endpoints-controller:0.2.1": unexpected status code [manifests 0.2.1]: 403 Forbidden
And if you go to look for the image you’ll see that it doesn’t exist. The work around is:
```bashgit clone https://github.com/jlewi/cloud-endpoints-controller.gitcd cloud-endpoints-controllergit checkout 0.2.1docker build . -t <YOUR DOCKER REGISTRY>/cloud-endpoints-controller:0.2.1docker push <YOUR DOCKER REGISTRY>/cloud-endpoints-controller:0.2.1```
This creates an image that the deployment can find until the official image is relaunched.
Credit: Moustafa Abdelhamid
Is Kubeflow in any way affected by the log4j security issue?
No. The only component in Kubeflow that uses log4j is an s3 connector which depends on JClouds which depends on log4j- however JClouds depends on log4j 1.x (like 90% of projects that depend on log4j) and the vulnerability only affected log4j 2.x, so rest easy you are safe.
How can I find more resources to help me with my New Year’s resolution to learn more about Kubeflow?
Arrikto’s events page has lots of content around upcoming webinars, meetups, and other Kubeflow related events. Definitely worth checking out.
December ‘21 Kubeflow Community Highlights
Slack
At the end of December ‘21 the Kubeflow Slack channel was 6,650+ members strong! (For those of you playing along at home, that’s an increase of almost 1,000 members since June ‘21!
GitHub
Here’s the Dec ‘21 GitHub stats for a few of the various Kubeflow projects.
- Kubeflow: 239 contributors, 11,046 stars, 37 new issues
- Pipelines: 142 new issues
- Katib: 14 new issues
- KServe: 37 new issues
- Manifests: 15 new issues
Meetups
It’s still too early for us to get together in person, but it doesn’t mean we can’t get together virtually to learn and discuss topics related to Kubeflow and MLOps! Arrikto has taken the initiative to set up a network of a dozen vendor neutral “Data Science, Machine Learning, MLOps and Kubeflow” Meetups in several major metros. Over 2,100+ community members have already signed up!
Join a Meetup near you and check out the calendar of upcoming talks.
Are you interested in speaking at a future Meetup?
Is your company interested in sponsoring a Meetup?
Would you like to be a co-organizer of a local Meetup? Send one of the organizers/hosts a message on Meetup.com!
Book a FREE Kubeflow and MLOps workshop
This FREE virtual workshop is designed with data scientists, machine learning developers, DevOps engineers and infrastructure operators in mind. The workshop covers basic and advanced topics related to Kubeflow, MiniKF, Rok, Katib and KFServing. In the workshop you’ll gain a solid understanding of how these components can work together to help you bring machine learning models to production faster. Click to schedule a workshop for your team.
About Arrikto
At Arrikto, we are active members of the Kubeflow community having made significant contributions to the latest 1.4 release. Our projects/products include:
- Kubeflow as a Service is the easiest way to get started with Kubeflow in minutes! It comes with a Free 7-day trial (no credit card required).
- Enterprise Kubeflow (EKF) is a complete machine learning operations platform that simplifies, accelerates, and secures the machine learning model development life cycle with Kubeflow.
- Rok is a data management solution for Kubeflow. Rok’s built-in Kubeflow integration simplifies operations and increases performance, while enabling data versioning, packaging, and secure sharing across teams and cloud boundaries.
- Kale, a workflow tool for Kubeflow, which orchestrates all of Kubeflow’s components seamlessly.