Bug 1985023

Summary:	[OVN SCALE] ovn-controller: Avoid fall back to recomputes when engine is aborted
Product:	Red Hat Enterprise Linux Fast Datapath	Reporter:	Numan Siddique <nusiddiq>
Component:	OVN	Assignee:	lorenzo bianconi <lorenzo.bianconi>
Status:	NEW ---	QA Contact:	Jianlin Shi <jishi>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	FDP 20.H	CC:	ctrautma, jiji, lorenzo.bianconi
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:		Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Numan Siddique 2021-07-22 16:49:44 UTC

Description of problem:

When ovn-controller is in the process of programming flows to ovs-vswitchd or
if there are any pending transactions, ovn-controller falls back to full recompute of logical flows in the next engine run.  This is the present design of ovn-controller.  But this full recompute can be very costly on very high scale environment.  ovn-controller should try to handle changes without falling back to full recompute.

One example is - Suppose a ct_zone engine needs to be full recomputed for any change, and if full recompute is not allowed then the engine is aborted and a full recompute of the whole engine is performed in the next run.  ct_zone engine full recompute ideally doesn't result in full recompute of openflows.


This needs to be improved in ovn-controller.


Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: