Bug 913606

Summary:	agent and server may go out of sync of avail if the server backfilled agents
Product:	[Other] RHQ Project	Reporter:	John Mazzitelli <mazz>
Component:	Agent, Core Server	Assignee:	John Mazzitelli <mazz>
Status:	ON_QA ---	QA Contact:
Severity:	high	Docs Contact:
Priority:	high
Version:	4.5	CC:	hrupp, jshaughn
Target Milestone:	GA
Target Release:	RHQ 4.6
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
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:
Bug Depends On:
Bug Blocks:	1094540

Description John Mazzitelli 2013-02-21 15:28:30 UTC

Here's the condition that I think created the problem:

1) Go in a steady state where agent is connected to the server and all avails are GREEN
2) Put server in MAINTENANCE mode
3) Wait for the server to backfill the agent
4) Put the server back in NORMAL mode

Notice the agent's platform never goes GREEN. This is because the agent doesn't know it has been backfilled and needs to tell the server about true state of avail.

Comment 1 John Mazzitelli 2013-02-21 21:24:46 UTC

gonna add a slight behavior change to the agent connect() message to the server. the server will reply back with a true or false - true if it has backfilled the agent already. If true, the agent will tell the plugin container that it needs to send a full avail report.

the server will also update the "last ping time" on the agent record to the current time when the connect came in - this will halt any future backfilling, at least for the following 5 minutes by default.

Comment 2 John Mazzitelli 2013-02-21 21:54:43 UTC

git commit to master: 277c8b1fc4379e6a652931b7db1bbdd4ce3ce5ac

Comment 3 Jay Shaughnessy 2014-04-04 17:44:40 UTC

I think this went into 4.6. Maybe 4.7.