Hi, can Notebook task run as a AWS Map Task? I tried this:

import os
import pathlib

from flytekitplugins.papermill import NotebookTask
from flytekitplugins.awsbatch import AWSBatchConfig

config = AWSBatchConfig(
    platformCapabilities="EC2",
)

backtest_wages = NotebookTask(
    name="test-notebook",
    notebook_path=os.path.join(
        pathlib.Path(__file__).parent.parent.absolute(), "notebooks/test-notebook.ipynb"
    ),
    render_deck=True,
    task_config=config
)

but got an error:

logger.debug(f"Starting _dispatch_execute for {task_def.name}")
AttributeError: 'function' object has no attribute 'name'

Hi! I'm trying to run a PoC using Flyte demo locally. The tasks are failing with OOM and I have already increased their mem limits to the current max allowed by the sandbox (1G). Is there a way to increase this overall limit for the sandbox configuration?

hello there, I have a bunch of print and logger.info messages in workflow, where do i see the outputs of those?

when I run the flyte console, It doesn't seem to print them

We’ve just seen a task failure with

containers with unready status: [fv6d6b4u4zgwuy-n10-0]|context deadline exceeded

- the task has not been retried (it would have three user retires set). Is this expected?

Hi everyone -- any idea when/if we will get a

1.6.2

backport for those of us stuck on Protobuf 3? Running into https://github.com/flyteorg/flytekit/pull/1657 on

1.2.13

Hi, I am following the guide to start a flyteconsole locally, but I always get

Cannot GET /console

response when trying to access http://localhost:3000/console . Wondering If I can get some hints on this. Thank you in advance!

hi team, when I run

make compile

and

./bin/flyte start --config flyte_local.yaml

to develop flyte locally, how can I specify the pod security context to allow container to run as root?

anyone have experience mounting s3fs inside a flyte task? Looks like I am getting fuse permissions issues

Hello, How do i configure the demo environment so that flyte will save the metadata and rawdata in my own s3 bucket? Is that possible? Thanks!

Apologies for spamming but here goes another question: I tried to use FrameProfilingRenderer but I got circular import error:

ImportError: cannot import name 'lazy_module' from partially initialized module 'flytekit' (most likely due to a circular import)

Any idea on how to fix this? Thanks!

Hi! Im running into an issue with returning a Pytorch model from a Task/Workflow (in demo environment). What am I doing wrong? This gets the Pod stuck in

Running

with logs `Failed to put data from /home/user/model.pt to s3://my-s3-bucket/data/.../model.pt; Original exception: Could not connect to the endpoint URL: \"http://flyte-sandbox-minio.flyte:9000/my-s3-bucket?list-type=2&prefix=data...`:

@task()
def train_model() -> bool:
  train() # model training, generates a .pt file in the container
  return FlyteFile(path="/home/user/model.pt")

@workflow
def wf():
    train_model = train_model()
    return train_model

This, just taking another random file in the same container completes without issues:

return FlyteFile(path="/home/user/.profile")

When I

exec

into the container, the

<http://model.pt|model.pt>

file is there.

i installed the flyteorg/flyte-binary v1.7.0 helm chart in my local k8s cluster, when enabled the auth, the flyte container can not started , the error is "{"json":{"src":"start.go:58"},"level":"fatal","msg":"Flyte native scheduler failed to start due to async future was canceled","ts":"2023-06-26T014846Z"}" how to deal with it??

hi team, i see multiple executions for the same task. is this normal? what are the different executions?

i used keycloak when sighed in ,it is forbidden

@Samhita Alla @David Espejo (he/him) please help me out

when sighed in using keycloak, it said bad gateway

Hey Where do we report bugs with flyte console? Looks like github issues are turned off for it The graph fails to expand half the time when using large sub workflows, our use case requires high concurrency so most of our flows are like this Anyone else had this problem?

We are running Flyte on GKE cluster, we use podTemplate to configure env vars and secrets. After updading the podTemplate, we were trying to run workflows using

pyflyte --image

but we’re getting this error (occasionally) -

Failed to apply default image tag "{{.image.demo.fqn}}:{{.image.demo.version}}": couldn't parse image reference "{{.image.demo.fqn}}:{{.image.demo.version}}": invalid reference format

Does anyone have any idea why it happens? Tnx

Hi, I am seeing this error when trying to register my workflow -

Failed with Unknown Exception <class 'AssertionError'> Reason: Cannot pass output from task n3 that produces no outputs to a downstream task

How do I find more information about this. Which task is this referring to?

When using

flytectl

, there are a few options around archiving/activating launchplans. I have been looking to move our wrapping code around launchplans to use

flytekit

, but I'm not seeing corresponding options. As far as I can tell, the

LaunchPlan.get_or_create

function just creates the object, but does not activate it. Am I missing something?

Hi, does anyone have a working example of using

custom_info

field within

TaskExecutionEvent

to relay custom information about a task run?

When trying to get the MPI Plugin running, I got this error:

│ flyte 2023/06/23 17:20:25 /go/pkg/mod/gorm.io/gorm@v1.24.1-0.20221019064659-5dd2bb482755/finisher_api.go:509                                                                                                                            │
│ flyte [1.083ms] [rows:1] SELECT count(*) FROM pg_indexes WHERE tablename = 'artifacts' AND indexname = 'artifacts_dataset_uuid_idx' AND schemaname = CURRENT_SCHEMA()                                                                   │
│ flyte {"json":{"src":"start.go:185"},"level":"panic","msg":"Failed to start Propeller, err: Error running FlytePropeller.: failed to load plugin - mpi: [PluginInitializationFailed] Error getting informer for %!s(\u003cnil\u003e), c │
│ flyte panic: (*logrus.Entry)

The MPI Kubernetes Operator works. I followed the documentation from here: https://docs.flyte.org/en/latest/deployment/plugins/k8s/index.html#deployment-plugin-setup-k8s and we deployed the most reason 1.7 flyte release.

When using ImageSpec, can we do the image building remotely? Can we launch builtkitd in a K8s pod and send our build requests there? Possibly with

envd context

Trying to create a channel #flyte-stripe in favor of the one in our workspace, but slack complains that name is already taken. I couldn't find this channel in browsing. Any chance Union team has already created a private channel under that name?

I am new to flyte and have been looking through the installation instructions and deployment paths. I am attempting to use the single cluster deployment path, but am trying to do so in an on-prem cluster instead of in the cloud. How do I need to adjust the helm chart(s) to cater for an on-prem deployment instead of cloud? Or is there documentation on the subject (I haven’t been able to find any)?

has anyone tried using deepspeed with flyte? found this example from ray docs, wondering if we can use the kuberay integration

Hi! I am just beginning to use Flyte and I would like to know if there is a way to inspect the results on a local workflow. Is it possible to visualize the project in this case?

Hi Community, I am new to Flyte and trying to run the

flytectl demo start

following the starter guide However I keep hitting error

Authenticated requests get a higher rate limit. Check out the documentation for more details.

And it doesn’t show what documentation I should follow (not in the starter guide) could you please help advise?

Hi, I was checking out

PythonTask

and

PythonFunctionTask

documentation under

flytekit.extend

, and am unsure about what the difference between the 2 are. Specifically, what does it mean by

PythonTask

should be used for task types when there is no python function to be executed?