Bug 1547374

Summary:	[Netvirt][NAT] After migration of NAPT switch there is no connectivity from an instance without FIP to external network
Product:	Red Hat OpenStack	Reporter:	Itzik Brown <itbrown>
Component:	opendaylight	Assignee:	Vishal Thapar <vthapar>
Status:	CLOSED WONTFIX	QA Contact:	Noam Manos <nmanos>
Severity:	high	Docs Contact:
Priority:	medium
Version:	12.0 (Pike)	CC:	aadam, asuryana, mkolesni, trozet, yma
Target Milestone:	z5	Keywords:	Triaged, ZStream
Target Release:	13.0 (Queens)
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	Netvirt
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2019-03-06 16:14:38 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:
Bug Depends On:
Bug Blocks:	1528948

Description Itzik Brown 2018-02-21 07:50:58 UTC

Description of problem:
After running the following scenario couple of times there is no connectivity 
from an instance without Floating IP to an external IP

Version-Release number of selected component (if applicable):
opendaylight-6.2.0-5.el7ost.noarch

How reproducible:


Steps to Reproduce:
1. Create an internal,external network, a router
2. Attach the network to the router
3. Launch an instance on a compute other than the one that holds the NAPT switch
4. Check connectivity from the instance to an external IP
5. Block the port 6653 on the node that holds the NAPT switch
6. Make sure that the switch is migrated
7. Check connectivity from the instance to an external IP
8. Delete all the resources above 
Verify that after couple of times the is no connectivity from the instance to an external IP.


Actual results:


Expected results:


Additional info:

Comment 1 Aswin Suryanarayanan 2018-02-21 07:53:50 UTC

There seems to be stale flows in the setup.  The packets should have hit the latter, but hit the former. Need to check if migration is leading to stale entries.

cookie=0x8000006, duration=28026.028s, table=36, n_packets=438, n_bytes=42924, priority=42,ip,tun_id=0x18716 actions=load:0x30e2c->OXM_OF_METADATA[0..23],resubmit(,44)
cookie=0x8000006, duration=25528.534s, table=36, n_packets=0, n_bytes=0, priority=10,ip,tun_id=0x18716 actions=load:0x30e2c->OXM_OF_METADATA[0..23],ct(table=46,zone=5003,nat)

Comment 2 Mike Kolesnik 2018-03-19 14:32:40 UTC

Does it still happen with the latest ODL version?

Comment 3 Itzik Brown 2018-04-15 10:20:22 UTC

It happens with:
opendaylight-8.0.0-5.el7ost.noarch

When the NAPT switch is moved to a controller it seems to be OK.
When it's moved to a compute node - it isn't.

Comment 9 Franck Baudin 2019-03-06 16:14:38 UTC

As per depreciation notice [1], closing this bug. Please reopen if relevant for RHOSP13, as this is the only version shipping ODL.

[1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/14/html-single/release_notes/index#deprecated_functionality

Comment 10 Franck Baudin 2019-03-06 16:17:10 UTC

As per depreciation notice [1], closing this bug. Please reopen if relevant for RHOSP13, as this is the only version shipping ODL.

[1] https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/14/html-single/release_notes/index#deprecated_functionality