What You’ve Been Asking Us Lately, December 2018 Edition
Can I use web applications to interact with streaming analytics?
Yes, for example you can use either REST or Websockets to publish data into a streaming project. Use Websockets to subscribe to pushed output from a project, or if you’re looking to have streaming output call an OData service, you can also use one of our OData output adapters.
HTTP Client JSON Output Adapter
HTTP Client XML Output Adapter
As a starting point for using REST to publish data to SDS, take a look at this tutorial:
Publish events to SDS via the REST interface
If you also need instructions on configuring the Streaming Web Service, see this tutorial:
Using the Streaming Web Service (SP11)
Also check out the Web Services Provider and Streaming Web Service sections in the documentation for additional information.
How can I fetch data from social media into streaming analytics?
Streaming analytics doesn’t currently include any pre-built adapters for social media API but you could easily build a streaming adapter for a social media source. You can start with this tutorial on building custom streaming adapters. This one show you how to build an MQTT adapter, but you can substitute the MQTT interface with the Twitter API.
Also, take a look at the Building Custom Adapters Guide for additional information.
Why can’t I start my project?
Issue: I created a project but when I try and start it, I get the following error:
When I look at the log file, I see the following:
Solution: The first potential issue here is the path you’re specifying for the CSV files in the adapter configuration. The filepaths need to be:
- Relative to the streaming server, and
- Located within the sandbox directory of the streaming server.
Since you are running streaming analytics, the server is on a Linux host. Even if you are using streaming with HANA Express (HXE), the server is running in a Linux virtual machine. The paths you are using indicate that you are asking the adapters to read from a directory on a Windows machine, which will not be accessible from the server.
This Getting Started Tutorial for Smart Data Streaming tutorial includes using a File/Hadoop CSV Input adapter which will help you get started.
The other potential issue is that the section of the log output that you included is for a Database Input adapter, but I don’t see one in the screenshot of your streaming project. Since there isn’t a Database Input adapter in your project, that section of the log output won’t be relevant to troubleshooting the current issue you are having.
Unable to successfully install streaming analytics in HANA Express
Issue: I have a physical HANA Express server and a HANA image and when try and install streaming analytics, the installation runs fine until I get this error message:
ALTER SYSTEM INITIALIZE SERVICE ‘streamingserver’ at ‘hxehost:39016’ with CREDENTIAL TYPE ‘PASSWORD’ USING ‘<password>’
And I get an error on the first try but when I execute it again, I get no error.
I can see the streaming service but it’s not started:
Open your streaming server trace file in $SAP_RETRIEVAL_PATH/trace/DB_HXE and look for this error:
ERROR – [LOCAL] [hxe] [3.7.4] Hazelcast CANNOT start on this node. No matching network interface found.
If you see this error, the issue is likely due to your /etc/hosts file mapping hxehost to 127.0.0.2 or something similar. Change that to a real IP # that can be resolved and bound to by the streaming server. After that, restart your HANA system.