normal ones

That's surprising

let me check with the team

running a spaceflights tutorial it seems to work okay - note that we use a special

_SharedMemoryDataSet

at runtime

whats the difference of a _SharedMemoryDataSet and a normal one ?

is that available or a custom dataset?

The parallel runner should use it automatically as far as I understand, so I'm not sure why this is popping up

I've asked the developers but I'm not sure when I'll get a response

do thinks work okay if you use SequentialRunner?

And if you're using Spark, the ParallelRunner will not work

sequential is working

It is failing in ParalelRunner in the _validate_catalog function

Have you explicitly declared MemoryDataSets in the catalog?

Updating from Kedro into SQL

yes

Ah so that may be causing the issue

so do they need to be declared there? MemoryDataSets are created implicitly if not present in the catalog

I created the entries through the cli

I can try to remove them

as long as the MemoryDataSets are used as outputs/inputs mid-pipeline they will be created by kedro without you declaring them

ok Thanks! I guess it should work without them explicitly written. I've just seen in the validate_catalog function it is explicitly checking for the presence of memory datasets. So if none is found it should work, but I have no idea how memory datasets are internally stored. Could they be overwritten or create some conflicts?

So if you scroll up and use the diagram I posted earlier - you don't have to declare

preprocessed_varieties

in the catalog, it will be produced by the first node and used by the

create_variety_table

. Kedro will create a MemoryDataSet at runtime to hand it between the nodes if it doesn't existing in the catalog.

I guess you might have found this already, but the docstring for

_validate_catalog

explains a bit what's going on here:

Ensure that all data sets are serializable and that we do not have non proxied memory data sets being used as outputs as their content not be synchronized across threads.

The second part about memory datasets is what's relevant here. As Joel said, default for parallel runner is that

_SharedMemoryDataset

is used rather than

MemoryDataSet

(see

ParallelRunner.create_default_data_set

for where this happens). In theory you could specify this dataset type explicitly in the catalog, but the fact that it's private means that's probably not a good idea, and I've never seen anyone do so. Just don't define them in the catalog and they will default to

_SharedMemoryDataset

and everything should work ok 🙂

Here's where it's all defined in case you're interested in what's going on under the hood: https://github.com/quantumblacklabs/kedro/blob/ded55eb824af25ea28ea9f5249317693a9b1574d/kedro/runner/parallel_runner.py#L26-L72. Just don't ask me how it works though since I've never actually looked at this code before 😄

Thanks a lot for your help @antony.milne @datajoely! now everything makes more sense! So as long as the datasets are pickable and not defined in the catalog everything should work fine. 🤞

Hello! I hope you could help me out with this doubt about the mlflow tracking server and how to set up everything in the mlflow.yml file of the kedro project. I have a database backend store in an external server to track metrics, etc and a SFTP server as artifact store for storing models in the same external server. In kedro-mlflow documentation I've seen I have to define the mlflow_tracking_uri variable but I'm not sure if a must write the sftp://user@host/path/to/directory of the artifact store or +://:@:/ of the database backend store. In my case I want to use both solutions and the mlflow.yml only give us one possible input. How can I set up both backend and artifact stores? Thanks!

Hello there, I hope you’re all doing well. Any guidelines/suggestions on how to deploy a kedro project to GCP / Cloud Composer. Should I upload the kedro project into the dag's folder? Or is there any other way to deploy it? Best,

Hello! I’m not sure we’ve had this come up before. Happy to help work through this - I wonder if any of the tutorials on the deployment guide apply here the same: https://kedro.readthedocs.io/en/stable/10_deployment/01_deployment_guide.html https://kedro.readthedocs.io/en/stable/03_tutorial/05_package_a_project.html

Thanks @User I'll dive into it and let the community know about any news (the good ones 🙂