Known issues

Limitations by region

You have these limitations if you provision Watson apps in the Frankfurt (Germany) or Tokyo (AP-North) IBM Cloud service regions.

Frankfurt region

Watson apps have these limitations in the Frankfurt region:

  • If you need a Spark runtime, you must use the Spark environment in Watson Studio for the model builder, modeler flow, and notebook editor tools.
  • The Batch Prediction and Real-time Streaming Predictions deployment types are not yet available.
  • Deep learning is not yet available. You can’t create deep learning notebooks or deep learning experiments.
  • The Neural Network Modeler is not yet available.
  • The Usage Statistics page for catalogs is not yet available.
  • The Total Catalog Assets information is not yet available.
  • Profiling of data assets that contain unstructured textual data is not yet available.
  • Cloud Functions are not yet available.

Tokyo region

Watson apps have these limitations in the Tokyo region:

  • If you need a Spark runtime, you must use the Spark environment in Watson Studio for the model builder, modeler flow, and notebook editor tools.
  • The Real-time Streaming Predictions deployment type is not yet available.
  • Deep learning is not yet available. You can’t create deep learning notebooks or deep learning experiments.
  • The Neural Network Modeler is not yet available.
  • The Total Catalog Assets information is not yet available.
  • Profiling of data assets that contain unstructured textual data is not yet available.
  • Multibyte characters are not supported in user-input fields in Data Refinery. Some fields allow multibyte characters but do not display them correctly.
  • Activity Tracker events for services provisioned in the Tokyo region are shown in the Activity Tracker service in the Sydney region.
  • Cloud Functions are not yet available.

Watson Knowledge Catalog

If you use the Watson Knowledge Catalog app, you might encounter these known issues and restrictions when you use catalogs.

Business glossary term associations

If you edit a policy or rule that was created before March 16, 2018, associations between the policy or rule and business terms that are used in the policy or rule are duplicated. When you view the business term, you’ll see the policy or rule listed twice on the Related Content page.

Refresh your browser after adding the Watson Knowledge Catalog app

If you add the Watson Knowledge Catalog app to your account using the Try-Out link on the landing page, you must refresh your browser to see catalog items.

Add collaborators with lowercase email addresses

When you add collaborators to the catalog, enter email addresses with all lowercase letters. Mixed-case email addresses are not supported.

Connections to Watson Analytics might not work

You might get an error when attempting to create a connection to Watson Analytics.

Object Storage connection restrictions

You cannot create data assets from Object Storage OpenStack Swift connections.

When you look at an Cloud Object Storage (S3 API) or Cloudant connection, the folder itself is listed as a child asset.

Multiple concurrent connection operations might fail

An error might be encountered when multiple users are running connection operations concurrently. The error message can vary.

Can’t enable data policies after catalog creation

You cannot enable the enforcement of data policies after you create a catalog. To apply data policies to the assets in a catalog, you must enable enforcement during catalog creation.

Data policies

If you use the Watson Knowledge Catalog app, you might encounter these known issues and restrictions when you use data policies.

Data is not anonymized in some project tools

When you add a connected data asset that has anonymized columns from a catalog to a project, the columns remain anonymized when you view the data and when you refine the data in the Data Refinery tool. Other tools in projects, however, do not preserve anonymization when accessing data through a connection. For example, when you load connected data in a notebook, you access the data through a direct connection and bypass anonymization.

Workaround To retain anonymization of connected data, create a new asset with Data Refinery:

  1. Open the connected data asset and click Refine. Data Refinery automatically includes the data anonymization steps in the Data Refinery flow that transforms the full data set into the new target asset.
  2. If necessary, adjust the target name and location.
  3. Click the Run button, and then click Save and Run. The new connected data asset is ready to use.
  4. Remove the original connected data asset from your project.

Business glossary terms need manual refresh

In Business Glossary, updates to business terms are not automatically refreshed when navigating back to the Business Glossary page with breadcrumbs. To refresh the Business Glossary page to view all updates, click the same button twice in the alphabet index.

Assets are blocked if evaluation fails

The following restrictions apply to data assets in a catalog with data policies enforced: File-based data assets that have a header can’t have duplicate column names, a period (.), or single quotation mark (‘) in a column name.

If evaluation fails, the asset is blocked to all users except the asset owner. All other users see an error message that the data asset cannot be viewed because evaluation failed and the asset is blocked.

Data Refinery

If you use Data Refinery, you might encounter these known issues and restrictions when you refine data.

CSV file limitations

Be sure that CSV files are correctly formatted and conform to the following rules:

  • Files can’t have some rows ending with null values and some columns containing values enclosed in double quotation marks.
  • Two consecutive commas in a row indicate an empty column.
  • If a row ends with a comma, an additional column is created.

Column name limitations

Be sure that column names conform to the following rules:

  • They are unique within the data set
  • They don’t contain any periods
  • They are not reserved words for R

System-level schemas aren’t filtered out

When creating a connection to IBM Db2 Warehouse on Cloud (previously named IBM dashDB), system-level schemas aren’t filtered out.

Target connection limitations

The following limitation applies to saving Data Refinery flow output (target data sets) to connections:

  • Data Refinery flows that have a target on a Compose for MySQL connected data asset are not supported.

Target file limitations

The following limitations apply if you save Data Refinery flow output (the target data set) to a file:

  • You can’t preview the file from the Data Refinery flow details page
  • Do not change the file format if the file is an existing data asset

White space characters are considered as part of the data

If your data includes columns that contain white space (blank) characters, Data Refinery considers those white space characters as part of the data, even though you can’t see them in the grid. Be aware that some database tools might pad character strings with white space characters to make all the data in a column the same length and this change will affect the results of Data Refinery operations that compare data.

“Replace missing values” operation does not show the replaced values

When you use the “Replace missing values” operation, the replaced values only show in the columns after you save and run the Data Refinery flow. View the Data Refinery flow output to see the replaced values.

t-SNE charts are not supported in Internet Explorer

The Data Refinery Visualization t-SNE chart is not supported in the Internet Explorer browser.

Watson Studio setup issues

You might encounter some of these issues when getting started with Watson Studio.

Can’t create assets in older accounts

If you’re working in an instance of IBM Watson that was activated before November, 2017, you might not be able to create analytical assets, like notebooks. If the Create button stays gray and disabled, you must add the Watson Studio app to your account. Click your avatar, click Add other apps, then add the Watson Studio app.

Error during login

You might get this error message while trying to log in to Watson Studio: “Access Manager WebSEAL could not complete your request due to an unexpected error.” Return to dataplatform.cloud.ibm.com and log in again. Usually the second login attempt works.

Manual installation of some tensor libraries is not supported

Some tensor flow libraries are preinstalled, but if you try to install additional tensor flow libraries yourself, you get an error.

Notebook issues

You might encounter some of these issues when working with notebooks.

Connection to notebook kernel is taking longer than expected after running a code cell

If you try to reconnect to the kernel and immediately run a code cell (or if the kernel reconnection happened during code execution), the notebook doesn’t reconnect to the kernel and no output is displayed for the code cell. You need to manually reconnect to the kernel by clicking Kernel > Reconnect. When the kernel is ready, you can try running the code cell again.

Using the predefined sqlContext object in multiple notebooks causes an error

You might receive an Apache Spark error if you use the predefined sqlContext object in multiple notebooks. Create a new sqlContext object for each notebook. See http://stackoverflow.com/questions/38117849/you-must-build-spark-with-hive-export-spark-hive-true/38118112#38118112.

Spark tasks might fail with missing AWS keys error

Respawning a failed executor during a job that reads or writes Parquet on S3 causes subsequent tasks to fail because of missing AWS keys.

Notebook kernel not started when opening a Scala notebook

You might notice that the notebook kernel is not running when you open a Scala notebook that uses Spark and custom Scala libraries. This situation arises when you use Scala libraries that are not compatible with the Spark version you are using, for example, if you use a Scala 2.10 jar file in a notebook with Spark 2.1.

To avoid this situation:

  1. Ensure that you use Scala 2.11 libraries with Spark 2.1.
  2. Run the following code in a Python notebook to remove the existing Scala libraries:
      !rm -rvf ~/data/libs/*
    
  1. Reload the libraries you need.

Connection failed message

If your kernel stops, your notebook is no longer automatically saved. To save it, click File > Save manually, and you should get a Notebook saved message in the kernel information area, which appears before the Spark version. If you get a message that the kernel failed, to reconnect your notebook to the kernel click Kernel > Reconnect. If nothing you do restarts the kernel and you can’t save the notebook, you can download it to save your changes by clicking File > Download as > Notebook (.ipynb). Then you need to create a new notebook based on your downloaded notebook file.

Connection to notebook kernel on Amazon EMR failed

If the notebook language, for example Python 3.5 with Spark isn’t displayed for the notebook, the notebook kernel couldn’t be started.

To verify that the Kernel Gateway to Amazon Elastic Map Reduce is started and its endpoints are accessible via the internet, run: curl https://<KG_EMR_URL>:<PORT>/api/kernelspecs -H "Authorization: token <Personal_access_token>"

The Kernel Gateway is accessible if a JSON list of the available kernels is returned. If not, you must reinstall Jupyter Kernel Gateway on Amazon EMR. For details, see Add an Amazon EMR Spark service.

Connecting to the notebook kernel on Amazon EMR is taking longer than expected

If your notebook kernel will not start, your Amazon Elastic Map Reduce service might have run out of Spark resources. You can free Spark resources by stopping the kernels of notebooks you aren’t using. Alternatively, you can stop all kernels by restarting the Kernel Gateway to the EMR cluster:

  1. Open the Amazon EMR console and log into the master node of the cluster.
  2. Enter wget https://raw.githubusercontent.com/IBMDataScience/kernelgateway-setup/master/install_kg_emr_bootstrap_script.sh to download the Kernel Gateway setup.
  3. Enter chmod +x install_kg_emr_bootstrap_script.sh to run the script.
  4. Enter ./install_kg_emr_bootstrap_script.sh --restart to restart the Kernel Gateway. You will be prompted for the port number.

Connection to Amazon EMR not available

If you keep running into problems connecting to Amazon Elastic Map Reduce, it is best you uninstall the Kernel Gateway and install it again:

  1. Open the Amazon EMR console and log into the master node of the cluster.
  2. Enter wget https://raw.githubusercontent.com/IBMDataScience/kernelgateway-setup/master/install_kg_emr_bootstrap_script.sh to download the Kernel Gateway setup.
  3. Enter chmod +x install_kg_emr_bootstrap_script.sh to run the script.
  4. Enter ./install_kg_emr_bootstrap_script.sh --uninstall to remove the Kernel Gateway.
  5. Enter ./install_kg_emr_bootstrap_script.sh to install the Kernel Gateway again.

Connection to IBM Analytics Engine service not available

The IBM Analytics Engine service instance that you selected to use for your notebook in Watson Studio might have been deleted or might not be running. Check if the service instance exists and is provisioned on the IBM Cloud Dashboard by clicking the navigation menu in Watson Studio and selecting Dashboard.

You can add a new IBM Analytics Engine service from your project’s Settings page in the associated services section.

Machine Learning issues

You might encounter some of these issues when working with IBM Watson Machine Learning components, including the Model Builder and Flow Editor.

Region requirements

You can only associate a Watson Machine Learning service instance with your Watson Studio project when the Watson Machine Learning service instance and the Watson Studio instance are located in the same region.

Flow Editor runtime restrictions

Watson Studio does not include SPSS functionality in Peru, Ecuador, Colombia and Venezuela.

Deployment issues

  • A deployment that is inactive (no scores) for a set time (24 hours for the free plan or 120 hours for a paid plan) is automatically hibernated. When a new scoring request is submitted, the deployment is reactivated and the score request is served. Expect a brief delay of 1 to 60 seconds for the first score request after activation, depending on the model framework.
  • For some frameworks, such as SPSS modeler, the first score request for a deployed model after hibernation might result in a 504 error. If this happens, submit the request again; subsequent requests should succeed.

AutoAI known limitations

  • Currently, AutoAI experiments do not support double-byte character sets.
  • If you stop an AutoAI experiment during a run, you will not be able to save pipelines that are generated. Currently, you can only save pipelines if the training run completes without interruption.
  • To interact programmatically with an AutoAI model, use the REST API instead of the Python client. The APIs for the Python client required to support AutoAI are not generally available at this time.

Streams flow issues

You might encounter some of these issues when working with a streams flow.

Streams flow restrictions

  • Event Streams and Kafka only display first ten rows in the preview of the Schema editor if no new messages are coming in.
  • Cloud Functions cannot load if your current Cloud Foundry organization is not set to your current location. To change your Cloud Foundry organization, go to the Profile and Settings page and select an organization within your current location.
  • If you restart the Streaming Analytics service, all existing stream flows are stopped.
  • A streams flow only works in projects that use IBM Cloud Object Storage.
  • The Lite plan for the Streaming Analytics service is good for 50 hours per month. If you exceed this limit, your streams flow cannot run.
  • Stream flows can be run only on the Streaming Analytics Container plans. Flows cannot run on the VM plans.
  • Support for the Microsoft Internet Explorer 11 browser is not yet complete.