pyTIgerGraph neighborLoader failed

I have a tigergraph instance hosted on GKE. Trying to utilize the GDS neighborLoader to fetch data for downstream tasks. Was able to install the neighbor_loader query. but iterating the loader to get batches took forever. Executing the installed query directly in GraghStudio gave me the following error

"kafkaError": "Failed, you need to upgrade to use Kafka streaming function.\nFailed, you need to upgrade to use Kafka streaming function.\n"

Did someone run into similar issues before?

Hi @kamadanen, do you have an enterprise license for Machine Learning Workbench? (Details here: Editions :: TigerGraph ML Workbench) If you do not have the enterprise version, you cannot use Kafka for data export. Additionally, you will need to provision your own Kafka cluster to facilitate the data export if you do have a license.

Hello @Parker_Erickson,

Thank you for your response! We (F5) are having a POC with TG and have a trial enterprise license that we have uploaded to GraghStudio. Is that sufficient? I do notice there is a mlwb binary that needs to be downloaded to activate the ML workbench. Since we got our trial license through email, we don’t have the credentials to login and download the enterprise mlwb. So I used the developer edition.

So for Kafka, you need the enterprise version. Let me check with our product manager/your TigerGraph sales person to see how we can enable you for the POC.

Thanks Parker, I think our sales person is Jason.

Hello @Parker_Erickson, I got my enterprise ML workbench license today. However, when I try to run the binary mlwb, I got the following question:

[8561] Error loading Python lib '/tmp/_MEIbusz1P/': dlopen: /lib/x86_64-linux-gnu/ version 'GLIBC_2.29' not found (required by /tmp/_MEIbusz1P/

It sounds like a glibc version issue, do you guys have a specification on software dependency for this binary to run?

Hmmm. Are you referring to the binary MLWB activator or the MLWB IDE? I don’t think we have a dependency for this binary.

@kamadanen , could you please share your sys info, like what OS you are on, and architecture (x86, arm, etc.)? Thanks!