Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1441651

Summary:	DVR L3 Agent Requires Two Restarts to Populate Removed Routes
Product:	Red Hat OpenStack	Reporter:	Benjamin Schmaus <bschmaus>
Component:	openstack-neutron	Assignee:	Assaf Muller <amuller>
Status:	CLOSED ERRATA	QA Contact:	GenadiC <gcheresh>
Severity:	high	Docs Contact:
Priority:	medium
Version:	10.0 (Newton)	CC:	amuller, bhaley, bschmaus, chrisw, jjoyce, nyechiel, srevivo, vaggarwa
Target Milestone:	z3	Keywords:	Triaged, ZStream
Target Release:	10.0 (Newton)
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:	openstack-neutron-9.3.1-2.el7ost	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2017-06-28 15:31:11 UTC	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 Benjamin Schmaus 2017-04-12 12:11:56 UTC

Description of problem:

Customers network team discovered that when the neutron fip-* namespace is deleted OR when the fg-* port is deleted from the br-int bridge, after restarting the 'neutron-l3-agent' service, the default gateway is missing within the fip-* namespace. It appears to be a race condition as a subsequent restart after the initial restart of the 'neutron-l3-agent' service adds the route back.


Version-Release number of selected component (if applicable):
OSP10

How reproducible:
100%

Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 3 Brian Haley 2017-04-12 13:09:41 UTC

I think I answered a similar bug on the rhos-tech list, here's my response from there:

There have been a couple of bugs fixed upstream (and backported) that might be related, the fact that adding the default route and arping fail make me think of this one:

(Race condition in update_gateway_port when two simultaneous router update occurs for the same router)
https://bugs.launchpad.net/neutron/+bug/1631513
https://review.openstack.org/#/c/385617/

Another related change is:

(After binding a floating IP to VM, the static route can't work in DVR)
https://bugs.launchpad.net/neutron/+bug/1571676
https://review.openstack.org/#/c/308068/

But I would think it's the first.

I do see that first change in the rhos-10.0-patches branch, so I'm assuming there is a build that has the change.  If that doesn't fix the problem then I'd need to investigate further.

Comment 4 Benjamin Schmaus 2017-04-12 13:22:31 UTC

Is that included in the rebase BZ:1430712

Comment 5 Brian Haley 2017-04-12 14:05:21 UTC

I don't see bug 1631513 listed in that BZ, but the fix for it was merged back in January, before some of the other bugs listed.

Comment 7 Benjamin Schmaus 2017-04-28 12:06:30 UTC

Customer is going to test applying the patches manually from the following:

https://review.openstack.org/#/c/385617/

https://review.openstack.org/#/c/413240/ 

I will report back the findings when they have completed the testing.

Comment 8 Benjamin Schmaus 2017-04-28 21:43:13 UTC

Assaf - Customer confirmed the manual edits fixed their issue.

Customer seems to be okay with waiting for rebase.  So I believe once the rebase comes out we can close this out as FIXED CURRENT RELEASE?

Comment 10 GenadiC 2017-06-14 15:48:46 UTC

Verified on openstack-neutron-9.3.1-2.el7ost.noarch
The code that solved the problem to the customer exists in this version

Comment 12 errata-xmlrpc 2017-06-28 15:31:11 UTC

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-2017:1594