Impala Connector MacOS

[Asimkz]

Hey everyone do we have a impala connector for MacOS? I can’t seem to get it to show up after installing the Big Data Connectors extension package.

[k10shetty1]

Hi Asimkz,

Did you try restarting your KNIME AP after installing the Big Data Connectors extension package?

If yes and the impala connector node is still not visible to you, can you try draging the node from here to your workflow in the KNIME AP ?

Thanks

[zyc]

Same problem here. Mac OS Ventura M1. Knime 4.7.

Pictures attached.

I’ve tried to drag the node too. Same problem again.

[mlauber71]

@zyc I have experienced problems with big data related extensions with Apple M1

Knime 4.7 - Apple Silicon - M1

@neiloxford I installed it on a M1 after first upgrading my ‚intel‘ KNIME version running thru Rosetta. The initial experience was that it is faster though I have not done that much since I experienced a problem with the Parquet file format (AP-19894). You could try to do a parallel installation and see for yourself.

[zyc]

So… I performed a fresh new instalation and start over.

First I thought is unusual see this component refering “Windows” on a Mac OS environment:

KNIME Extension for Big Data Commons on Windows

After install the Impala Connector, and after restarting the Knime, I’ve took a look at the error logs. There were a lot of error messages.

Following the stack is possible to find the probably unlinked dependency.

The org.apache.hadoop.client.

I’ve tried to solve the broken dependency installing others extensions but I got the same error.

Could someone suggest something to solve that?

Pictures attached.

[mlauber71]

@zyc I think KNIME currently has a problem with apache/parquet extension/formats on Apple Silicon machines (AP-19894) and will have to fix that. Until then you might have to stick to the ‘intel’ version and use it thru Rosetta (MacOSX will do that automatically you will just have to install it). I think this is very unfortunate and I am hoping for a fix since December since the new M1/M2 chips seem to offer a significant increase in speed (20% or so).

[zyc]

Thanks @mlauber71. So I’ll run the Knime under Rosetta until the M1 version get fixed.

[zyc]

I decided to install the Knime over Homebrew this time running.

$ brew install --cask knime

knime

Now the Intel version was installed instead of M1 as you can see:

Now, the Impala Connector is been shown up in Node Repository section.

[sascha.wolke]

Hi,

We released a fix with KNIME 4.7.2 that should fix the Big Data installation on Arm64 Mac.

@zyc does this resolve the Problem with the Impala Connector?

Cheers,
Sascha

[mlauber71]

@sascha.wolke at least my problems with the Big Data formats and Apple silicon seem to be gone and I can enjoy the 20% speed increase with the native M1/M2 version of KNIME.