How to isolate and replicate a TDM or a TDM on Spark issue

[TalendSolutionExpert]

Overview

Below are some steps that could help to isolate or replicate a TDM issue or a TDM on Spark issue.

Description

• Check the last change done before break
  • Find out what was changed in the Route/Job before issue surfaced
  • Where was the change made – Structure/Mapping?
  • What is after an upgrade/migration?
  • Did the input document / format change?

• Scope of the issue: What works or does not work
  • Is it a user/Studio/Machine specific issue?
  • Is it Mapping/Structure related?
  • Does it depends on a specific representation / data used?
  • Size of input (in case Spark is not used)
  • Connection to the cluster: check if working fine without TDM / TDM on Spark components
  • Performance related issue – on Studio/Jobserver/Runtime/Cluster

• Working/non-working scenarios
  • Find out if there is a specific input/use case that always fails
  • Determine simplest case where problem occurs (process of elimination)

• Steps to reproduce
  • Exact environment details (Talend version, Cluster details…)
  • Create a mockup/sample Structure/Map/Route/Job in-house
  • Customer data: get sample input/output data if it is data dependent issue
  • Get the entire Job/Route (and in some cases, the project or workspace) for cases where issue cannot be easily replicated