Error compute kernel failure new Statistics Feature

Jailbreaker

Hello,

I updated our Dataiku Plateform from 5.X to the 8. version.

So when we try to use the new Statistics feature on a Dataset, I get the following error:

Unexpected EDA kernel failure

com.dataiku.dip.exceptions.CodedException: EDA docker kernel failed to start, caused by: CodedException: EDA docker kernel failed to start, caused by: SecretKernelTimeoutException: Subprocess failed to connect, it probably crashed at startup. Check the logs., caused by: SocketException: Socket closed

Here are the logs I got from the backend.log file

[2020/11/25-19:17:01.508] [qtp1528769018-365630] [DEBUG] [dku.tracing]  - [ct: 1] Start call: /api/futures/get-update [GET] user=admin [futureId=1EMRVvBF]
[2020/11/25-19:17:01.508] [qtp1528769018-365630] [DEBUG] [dku.tracing]  - [ct: 1] Done call: /api/futures/get-update [GET] time=1ms user=admin [futureId=1EMRVvBF] thrown=ExecutionException
[2020/11/25-19:17:01.508] [qtp1528769018-365630] [ERROR] [dip.controllers]  - API call '/dip/api/futures/get-update' failed
java.util.concurrent.ExecutionException: com.dataiku.dip.exceptions.CodedException: EDA local kernel failed to start
        at java.util.concurrent.FutureTask.report(FutureTask.java:122)
        at java.util.concurrent.FutureTask.get(FutureTask.java:192)
        at com.dataiku.dip.eda.compute.engine.LazyRunner.getOrBuildRunner(LazyRunner.java:88)
        at com.dataiku.dip.eda.compute.engine.LazyRunner.compute(LazyRunner.java:38)
        at com.dataiku.dip.eda.compute.engine.SmoothlyClosableRunner.compute(SmoothlyClosableRunner.java:23)
        at com.dataiku.dip.eda.compute.caching.CachedRunner.compute(CachedRunner.java:83)
        at com.dataiku.dip.eda.worksheets.CardComputeService.computeResultsWithRunner(CardComputeService.java:78)
        at com.dataiku.dip.eda.worksheets.CardComputeService.computeResults_NT(CardComputeService.java:67)
        at com.dataiku.dip.eda.worksheets.CardComputeService$1.compute(CardComputeService.java:50)
        at com.dataiku.dip.eda.worksheets.CardComputeService$1.compute(CardComputeService.java:35)
        at com.dataiku.dip.futures.SimpleFutureThread.execute(SimpleFutureThread.java:36)
        at com.dataiku.dip.futures.FutureThreadBase.run(FutureThreadBase.java:88)
Caused by: com.dataiku.dip.exceptions.CodedException: EDA local kernel failed to start
        at com.dataiku.dip.eda.compute.engine.python.LocalPythonKernel.start(LocalPythonKernel.java:33)
        at com.dataiku.dip.eda.compute.ComputeService.createSampleKernel(ComputeService.java:146)
        at com.dataiku.dip.eda.compute.ComputeService.access$000(ComputeService.java:52)
        at com.dataiku.dip.eda.compute.ComputeService$1.call(ComputeService.java:93)
        at com.dataiku.dip.eda.compute.ComputeService$1.call(ComputeService.java:87)
        at java.util.concurrent.FutureTask.run(FutureTask.java:266)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
Caused by: com.dataiku.dip.io.SocketBlockLink$SecretKernelTimeoutException: Subprocess failed to connect, it probably crashed at startup. Check the logs.
        at com.dataiku.dip.io.SocketBlockLink.waitForConnection(SocketBlockLink.java:70)
        at com.dataiku.dip.io.SecretProtectedKernelLink.waitForProcess(SecretProtectedKernelLink.java:39)
        at com.dataiku.dip.io.SecretProtectedKernelLink.waitForProcess(SecretProtectedKernelLink.java:48)
        at com.dataiku.dip.io.PythonSecretProtectedKernel.start(PythonSecretProtectedKernel.java:108)
        at com.dataiku.dip.eda.compute.engine.python.LocalPythonKernel.start(LocalPythonKernel.java:29)
        ... 8 more
Caused by: java.net.SocketException: Socket closed
        at java.net.PlainSocketImpl.socketAccept(Native Method)
        at java.net.AbstractPlainSocketImpl.accept(AbstractPlainSocketImpl.java:409)
        at java.net.ServerSocket.implAccept(ServerSocket.java:545)
        at java.net.ServerSocket.accept(ServerSocket.java:513)
        at com.dataiku.dip.io.SocketBlockLink.waitForConnection(SocketBlockLink.java:66)
        ... 12 more
[2020/11/25-19:17:01.509] [qtp1528769018-365630] [INFO] [dip.controllers]  - Err for /dip/api/futures/get-update

Jailbreaker

I finally install the missing package directly from the command line, everything is going fine now! The command I run is

dss_data/bin/pip install sortedcontainers

AdrienL

Attached backend logs seem to show that the sortedcontainer package is missing. You can try rebuilding your python built-in environment.

Jailbreaker

I ran the rebuilt command and added this librairy, restart my DSS Instance but still the same error.

Attached a screenshot our installed packages in the python_env

AdrienL

The same error as in the backend logs (which was the one I was addressing, about a *local* EDA kernel) or the same error as in you original post's main text (about a *docker* EDA kernel)?
If it's the latter, try unsetting docker configuration and using a local kernel to check whether this works.

Jailbreaker

Even with the "local" Kernel, I get the same error

Unexpected EDA kernel failure

com.dataiku.dip.exceptions.CodedException: EDA local kernel failed to start, caused by: CodedException: EDA local kernel failed to start, caused by: SecretKernelTimeoutException: Subprocess failed to connect, it probably crashed at startup. Check the logs., caused by: SocketException: Socket closed

AdrienL

Yes, this is the same overall error, the details are in the backend logs you attached the first time, please attache those again.

Jailbreaker

It's the same error I think, as you can see in the attached log file

AdrienL

Strange. Can you try, in a local notebook in the built-in environment, to `import sortedcontainer`?

Jailbreaker

Seems that I add it in the python_env, not in the built in Environment!

Because is the Jupyter Notebook, by default, it doesn't import it. But when I change the kernel to my python_environment, all is OK.

What I want is to add this package 'sortedcontainers' to my built in environement!

So how can I add it to the built in Environment so?

AdrienL

I am not sure I understand what you mean. I think the easiest would be to share an instance diagnostic with support@dataiku.com (from Administration > Maintenance > Diagnostic tool). If the resulting file is too large for mail (> 15 MB), you can use https://dl.dataiku.com to send it. If so, please don't forget to send the link that is generated when you upload the file.

Jailbreaker

OK

AdrienL

You should send that to support@dataiku.com (or on support.dataiku.com), as these files can contain things you may not want public, and support is better armed than community to answer on the specific configuration that you have and that is not common to other users.