Bug 113885

Summary:	permissions denorm triggers slow on postgres
Product:	[Retired] Red Hat Enterprise CMS	Reporter:	Bryan Che <bche>
Component:	other	Assignee:	Aram Kananov <akananov>
Status:	CLOSED WONTFIX	QA Contact:	Jon Orris <jorris>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	nightly
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2006-09-05 17:51:16 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Bryan Che 2004-01-19 21:46:35 UTC

I've been populating a cms postgres server, and the data population is
much slower now than before.  With around 5,000 published articles so
far, creating and publishing each additional article is taking around
5 minutes.  Previously, the slowest performance I saw populating on a
heavily used postgres db was around at most 1 minute/article.

The biggest problem seems to be doing operations like

update object_context
set context_id = 7149
where object_context.object_id = 15004

After turning on stats gathering for postgres, I see lots of
operations like these during population, and they each take up to
around 12 seconds to complete.  I think the culprit is the
object_context_dnm_trg trigger on object_context.  

I am currently populating an oracle server as well with statspack
enabled, but it's too early to know yet if oracle has this problem too.

Comment 1 Bryan Che 2004-01-20 18:14:40 UTC

just an update on oracle performance for population with the triggers.
 Creating and publishing an article on oracle took at most around 3
seconds at the same changelist that postgres is taking around 5 minutes.

Comment 2 Aram Kananov 2004-01-20 19:16:59 UTC

p4 # 39541 should improve performance on pg

Comment 3 Bryan Che 2004-01-21 15:14:45 UTC

Now, populating articles is going at a rate of about 50
seconds/article.   This is a big improvement but still lags Oracle by
quite a bit.

Comment 4 Jon Orris 2006-09-05 17:51:16 UTC

Closing old tickets