Skip to main content

Performance degradation when going on 7.1.1

  • November 14, 2014
  • 4 replies
  • 22 views

Luc_Lecocq
Forum|alt.badge.img+2
Hi,

I was previously running 7.1.0 SP2, and upgraded to 7.1.1 on a 4 nodes cluster
Prior to this upgrade, I was running regular queries every 15mn, the database was loaded continuously with new data, projections had been created manually in order to optimize query execution.
The execution times for these queries was relatively stable.
Right after the upgrade to 7.1.1 and restarting the database, I observed an increase in the execution time of most queries, of almost 60% (from 250 sec to 400 sec).
Is there any reason that could explain this ?
I want to be clear: the only thing I did was: stop DB, update vertica, restart db, strictly no change in incoming data flow or config or data model.

4 replies

Akshar_Dave
  • New Participant
  • December 12, 2014
This is concerning. Hopefully community shares  here if they ran into similar issues after upgrade.

Sashidhar_Arcot
Forum|alt.badge.img
  • Participating Frequently
  • December 27, 2014
Some settings like IO scheduler or Transparent Huge pages could have changed. Please make sure to have met all the pre-requisistes.

Open a support ticket if this still bothers you and we will look into it. 

id10t
Forum|alt.badge.img
  • Participating Frequently
  • December 27, 2014
Hi!

Try next keywords: SAL, reindex, CRC tool, vertica upgrade + SAL reindex.

id10t
Forum|alt.badge.img
  • Participating Frequently
  • December 28, 2014
Hi!

May be resource pools config changed?

I will summarize:
  • IO scheduler
  • Huge Pages
  • read ahead value (changed after upgrade on lower value)
  • resource pool
  • reindex (SAL)