DBR 15.4 LTS Compatibility Confirmed with Zingg 0.4.0 in Initial Testing

Luke B.

·Nov 21, 2024 02:48 PM

Heads up: It looks like DBR 15.4 LTS is compatible with zingg 0.4.0 ... at least in some early testing.

20 comments

· Sorted by Oldest

Sonal G.
·
We are seeing some issues with 15.4 but it worked on 16.0 ML Luke B.
Renzo S.
·
hello there Sonal G. , I'm Renzo and I work for Tranzact can we know what were these issues? because we're currently using zingg in Databricks, but we've encountered an issue, after following this guide FebrlExample.ipynb, related to this message:
Py4JJavaError: An error occurred while calling o521.execute. : zingg.common.client.ZinggClientException: org.apache.spark.sql.types.StringType$; local class incompatible: stream classdesc serialVersionUID = 3796071416192072411, local class serialVersionUID = 1779832429914676547
we'd like to know if this is a bug or a misconfiguration
Sonal G.
·
Which phase are you running?
Renzo S.
·
it's the link phase
Sonal G.
·
And at which line is this error coming?

Renzo S.

·

link_options = ClientOptions([ClientOptions.PHASE, 'link'])
link = ZinggWithSpark(args, link_options)

# clear old outputs from any prior runs
dbutils.fs.rm(linked_output_path, recurse=True)

link.init()

link.execute()

Renzo S.
·
at the last, execute
Sonal G.
·
Do you have a prebuilt model that you are using? Seems like a class version compatibility problem
Renzo S.
·
yes, we do
Sonal G.
·
In that case it is better to run train first so that the version of the model is the same
Renzo S.
·
I'm going to try this, thanks!
Renzo S.
·
hello Sonal, good evening unfortunately, retraining the model didn't work
Renzo S.
·
we're using DBR 15.4 LTS with zingg 0.4.0, btw
Sonal G.
·
And dbr 16.0 with ml?
Renzo S.
·
I'll try that, thank you