Problem or anti-pattern workloads
You must be aware of the guidelines for moving antipattern or problematic workloads.
Follow these guildelines to move problem or anti-pattern workloads.
- Select workloads that are complex and are complaining about impala performance, queuing issues, pool sizes, insufficient memory limits, scanning random datasets.
- Use WXM to identify the workloads that miss SLAs in WXM, that are killed due to running out of memory, that are running slow and would benefit from data caching.
- Consider using Hue or Impala Shell or some thick client JDBC tool like Dbeaver or Squirrel
SQL for JDBC workloads.
- You must be aware of the standard JDBC compatibilty issue for thick clients.
- Segregate each problem workload by assigning it to its own VW.
- Redirect BI tool to point to the VW for these workloads.
- A large sized VW is recommended for these workloads to perform better in CDW PvC 1.3.1 than on CDP Private Cloud base.
- To optimize performance, set a high value for
EXEC_TIME_LIMIT_S.