https://flyte.org
Join Slack
Hello, I'm going to deploy my workflow on the local container. But the workflow fails with this mess...
# flyte-support
l
Hello, I'm going to deploy my workflow on the local container. But the workflow fails with this message:
Copy code
[1/1] currentAttempt done. Last Error: USER::containers with unready status: [fba452891990446a7a49-n0-0]|Back-off pulling image "cosine:2ce90d3ad559b4e5b1981af8726f4d19eeedc835"
Is there any way I can use to debug what's happened?
👀 1
f
It seems you did not push the container to a registry
So one option when working with demo / sandbox is to build the container into demo
This is done by mounting the source inside - handled automatically by flytectl
Let me share the docs
But I think in a doc restructure we missed this
We will add it to getting started are 2
Custom images. Cc @echoing-queen-96254 @billowy-sundown-31926
👀 1
✅ 1
l
I've built the image using 
docker_build_and_tag.sh
, and packed the workflow using 
pyflyte --pkgs flyte.workflows package --image <registry/repo:version>
. But I don't know how to run my workflows using this image, executing 
flytectl create execution --project flytesnacks --domain development --execFile exec_spec.yaml
results in the same container with unready status error
I finally managed to deploy my workflow to local demo container. I noticed that when the task passes a type that flyte doesn't support, the object is converted using 
PythonPickle
, and attached with a s3 url. Does it mean the dumped file is uploaded to aws?
f
In local case to the bundled minio - which has an s3 like interface
But you are right in prod it will be one of the blob stores
S3, gcs, abfs, … this is extensible
l
Yeah, we'll use our private data storage service. I found this doc: https://docs.flyte.org/projects/flytekit/en/latest/data.extend.html but it's somewhat.. simple. I still don't know exactly how to extend flyte to use my custom data storage
f
@lemon-minister-88597 let me dm you I can definitely help with this. Can we do a quick VC call and we can discuss
l
It seems that we have jet lag here.. I'm going to have some meetings in next few hours. Let me have a try before we talk :D
f
Sure
Ya I am in PDT, where are you at
397 Views
Open in Slack
PreviousNext
Powered by

Flyte

Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Powered by