Working with StreamSets Support

This is a guide to help you, our customers, gain insight into how to troubleshoot issues with StreamSets products and to help you more effectively engage your StreamSets Support team. We hope that these recommendations help to streamline the process of working with StreamSets Support and reduce the amount of time spent with us when investigating and resolving issues.

Note: The use of StreamSets Support resources requires that you have an active Enterprise License subscription for StreamSets products. If you do not have an Enterprise License, you can join our community Slack channel or ask us at ask.streamsets.com for assistance with your issues.

Contacting StreamSets Support for Assistance

  1. Always start by opening a support ticket with us through streamsets.zendesk.com
  2. Follow the steps in this guide to gather the necessary data for us to begin troubleshooting your issue
  3. If there is a need to request immediate engagement on an existing support ticket, please call our Support Hotline: 1(888)366-0699
  4. For P1 support tickets opened during the weekend (Saturday/Sunday), please make sure to call the Support Hotline after opening the ticket.

Information to Provide When Opening a Ticket

When opening a support ticket in our support portal, it will speed up the process of troubleshooting if you can provide as much information as possible up front, which includes the following:

  • Detailed explanation of the issue:
    • Type of issue (error message(s) reported, functionality not working as expected, performance problem, OOM issue, etc.)
    • The pipeline/job export if the issue is related to a stage, a pipeline, or a job
    • Stack trace of the error message
    • Is it a new pipeline/job? Does it affect only one pipeline/job or is the impact more widespread?
    • Has the issue occurred before?
    • Is the issue reproducible? What steps lead to this issue?
  • Product version: StreamSets Data Collector, StreamSets Control Hub, or StreamSets Data Protector
  • Environment type: production or non-production
  • Priority of the issue (for priority definitions: see our Support Policy)
  • Data for troubleshooting

Data for Troubleshooting

To help us start working on the issue as soon as possible, we will need you to provide relevant data, which includes the following:

  • Logs (sdc.log or dpm.log)
  • Pipeline or Job export
  • Environment and configuration information of the product

The requested files may differ based on which product you are using and what issue you are seeing:

  • StreamSets Data Collector: Support Bundle
  • StreamSets Control Hub + Dataflow Performance Manager: dpm.log (if On-Prem installation), configuration, print screens which clearly describe the issue
  • StreamSets Data Protector: Support Bundle (if the issue is happening on the Data Collector side) or dpm.log (if the issues is happening on the Control Hub side)

StreamSets Data Collector Logs - Support Bundle

To make it convenient for our customers to provide log files, our StreamSets Data Collector product has an built-in feature – Support Bundle. A support bundle is a ZIP file that includes Data Collector logs, environment and configuration information, pipeline JSON files, resource files, and pipeline snapshots. A user can either upload the bundle directly to StreamSets or download it for manual upload to the support ticket.

What benefits does the Support Bundle provide?
  • Easiest way to send us log data
  • Choosing what data (based on chosen generators) you want to share with us
  • Auto-redaction of all passwords entered in pipelines, configuration files, or resource files
  • Ability to create custom rules to redact additional sensitive data, e.g. hostnames (see Customizing Generators)

How to Generate a Support Bundle

Before you start, please double check that you already have a customer.id file installed in the $SDC_DIST/data directory. If you are missing this file, you can still upload the Support Bundle, however, copy paste the SDC ID from the SDC UI and send it along with the description of the issue. At your earliest convenience, install the customer.id after the Support team provides you one. The customer.id will help our team to find the respective bundle and identify it quickly.

Note on pipeline names: The support bundle generates a ZIP file which may contain a folder with all pipelines. Therefore if you have an issue with a respective pipeline, mention the pipeline ID in the description of the issue when creating your support ticket.

Memory Issues

If you are seeing memory issues, we may need to analyze a heap dump that you will need to provide. By default, when Data Collector encounters an out of memory error (OOME), it creates a heap dump. By default, heap dump files are written to the file defined in the SDC_LOG environment variable and use a naming convention that allows generation of multiple heap dump files with the following format:

$SDC_LOG/sdc_heapdump_${timestamp}.hprof.

In case the heap dump was not generated as a result of the memory issue or if the environmental variable is set to “false”, it is very important to generate the heap dump at runtime when the SDC is experiencing the memory issue. You can use one of the following commands to generate the heap dump in realtime:

sudo -u sdc jcmd <pid> GC.heap_dump <filename>.hprof

This is equivalent to the old jmap command:

sudo -u sdc jmap -dump:file=<filename>.hprof <pid>
  • For more information on configuring heap dump creation, see our documentation here

Detailed reference information on use of the jcmd command is available from Oracle’s Java reference documentation here

Performance Issues

For performance issues that can be described as “The UI is hanging and not responding” or “The pipeline is stuck in the starting/stopping state”, we usually need to look into thread dumps. Because a thread dump is a snapshot of threads from the moment when it is generated, we usually ask for several (5) thread dump files generated 5 seconds apart. Multiple thread dumps in succession give us a better overview of threads that could be problematic. You can generate a thread dump by running one of the following commands:

sudo -u sdc jstack -l <pid> > thread_dump.txt

(-F option (only if pid is unresponsive) -F Force a stack dump when ‘jstack [-l] pid' does not respond)

sudo -u sdc jcmd <pid> Thread.print > thread_dump.txt

Uploading Large Files through a Support Ticket

Zendesk, our support ticket management system vendor, has a 20 MB limitation for ticket attachments by default. StreamSets Support has solved this problem by leveraging SendSafely, which provides for a seamless drop-in replacement for Zendesk’s built-in file attachment system. Customers should be able to upload files of unlimited size through a support ticket attachment.

However, if you experience an issue with uploading large files, please let us know in the support ticket and we will work with you to resolve the issue.

Receive Updates

Receive Updates

Join our mailing list to receive the latest news from StreamSets.

You have Successfully Subscribed!