Bug 1384696 – Node sends Ready Event before network is configured

Bugzilla will be upgraded to version 5.0. The upgrade date is tentatively scheduled for 2 December 2018, pending final testing and feedback.

Bug 1384696 - Node sends Ready Event before network is configured

Summary:	Node sends Ready Event before network is configured

Status:	CLOSED ERRATA

Aliases:	None

Product:	OpenShift Container Platform
Classification:	Red Hat
Component:	Networking (Show other bugs)
Sub Component:
Version:	3.3.0
Hardware:	Unspecified Unspecified

Priority	high Severity medium
Target Milestone:	---
Target Release:	---
Assigned To:	Dan Williams
QA Contact:	Meng Bo
Docs Contact:

URL:
Whiteboard:
Keywords:

Depends On:
Blocks:
	Show dependency tree / graph

Reported:	2016-10-13 17:49 EDT by Eric Jones
Modified:	2017-03-08 13 EST (History)
CC List:	6 users (show)

See Also:
Fixed In Version:
Doc Type:	Bug Fix
Doc Text:	Previously nodes in an Openshift cluster using openshift-sdn would occasionally report readiness and be assigned pods before networking was fully configured. Nodes now only report readiness after networking is fully configured.
Story Points:	---
Clone Of:
Environment:
Last Closed:	2017-01-18 07:42:44 EST
Type:	Bug
Regression:	---
Mount Type:	---
Documentation:	---
CRM:
Verified Versions:
Category:	---
oVirt Team:	---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---

Attachments	(Terms of Use)
Add an attachment (proposed patch, testcase, etc.)

External Trackers
Tracker	ID	Priority	Status	Summary	Last Updated
Red Hat Product Errata	RHBA-2017:0066	normal	SHIPPED_LIVE	Red Hat OpenShift Container Platform 3.4 RPM Release Advisory	2017-01-18 12:23:26 EST

Groups: None (edit)

Description Eric Jones 2016-10-13 17:49:10 EDT

Description of problem:
A node can send a Ready Event to the master (indicating it is prepared to receive pods) before the network is completely configured. In the specific instance that we found this issue the nodes did not have the networkConfig value set int he node-config.yaml but the nodes were seen as "Ready" even though they were not.

Version-Release number of selected component (if applicable):
OpenShift Container Platform 3.3

Comment 2 Ben Bennett 2016-10-26 15:52:04 EDT

Does this happen post-CNI?

Comment 3 Dan Williams 2016-10-27 11:52:43 EDT

This should no longer happen post-CNI, due to the new bits in the network plugin Status() hook that will return a periodic error to kubelet until the node has a HostSubnet assigned.

Comment 4 Troy Dawson 2016-11-02 13:52:05 EDT

What is post-CNI?
Or more important, is this fixed in OCP 3.3.1?  or OCP 3.4.0?  I'm trying to figure out which errata to attach it to.

Comment 5 Ben Bennett 2016-11-02 14:51:48 EDT

We changed the OpenShift networking plugin to use the CNI interface that changed quite a few of the internals for the better.

That happened in 3.4.  So it will not be fixed in 3.3.1.

Comment 6 Troy Dawson 2016-11-02 17:46:40 EDT

Thank you for the explanation.

Comment 8 hongli 2016-11-03 06:29:13 EDT

verified on OCP 3.4.0.19 and con not reproduce the issue.

Will keep an eye on it since some CNI fixes have not been merged to this build.

Comment 10 errata-xmlrpc 2017-01-18 07:42:44 EST

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2017:0066

Note You need to log in before you can comment on or make changes to this bug.