Sunday, July 7, 2019

Leveraging Machine Learning to diagnose Real Application Clusters Issues & Incidents

Today's post is a shout out to a much needed presentation by Anil Nair Oracle Senior Principal Product Manager on Leveraging Machine Learning to diagnose Real Application Clusters Issues & Incidents.

Here is a brief summary of the presentation by Anil:

  • Scalability without changing Application Code.
  • Problem Resolution Paths & Approaches.
  • Reactive & Proactive Diagnosis.
  • Common Problems such as Hangs, Performance Issues, Deadlocks etc.
  • Case Studies of Incidents & Issues in Real Application Clusters (RAC).
  • Automatic Notification & Diagnosis Collection with Trace File Analyzer (TFA).
  • Cluster Health Advisor TFA SMTP Notifications.
  • Configuring OraChk to run in DAEMON mode.
  • Node Evictions due to Memory Pressure.
  • RAC Misconfigurations.
  • Oracle Memory Guard.
  • Aggregated Data by Process Type provided by CHM.
  • Sample Problems & Resolutions.
  • Reconfiguration Diagnosability.
  • Dynamic Remastering (DRM) Diagnosability.
  • Grid Infrastructure Management Repository (GIMR).
  • Autonomous Health Framework (AHF) collects much of the data that OSWatcher collects.
  • Documents & Resources
  • Whats New?
  • Oracle Autonomous Health Framework (New).
  • Applied Machine Learning Diagnostics (New).
  • Autonomous Health - Database Performance (New).
  • Cluster Health Advisor Graphical (CHAG) (New).
As should be clear by this excellent presentation, Leveraging Machine Learning to diagnose Real Application Clusters Issues & Incidents makes the whole RAC diagnosis process easier & faster.

Cheers.