Bug 2263502

Summary: [17.1][OVN][DVR][SNAT] tcp communication between vms on different platforms using provider network is broken when snat in enabled
Product: Red Hat OpenStack Reporter: Flavio Piccioni <fpiccion>
Component: rhosp-openvswitchAssignee: Miro Tomaska <mtomaska>
Status: CLOSED CURRENTRELEASE QA Contact: Bharath M V <bmv>
Severity: high Docs Contact:
Priority: high    
Version: 17.1 (Wallaby)CC: bmv, chrisw, gregraka, ihrachys, jlibosva, ltamagno, mariel, mblue, mburns, mtomaska, scohen
Target Milestone: z4Keywords: TestOnly, Triaged
Target Release: 17.1Flags: mtomaska: needinfo-
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: rhosp-openvswitch-3.3-1.el9ost Doc Type: Bug Fix
Doc Text:
Before this update, links were failing for instances connected directly to an external network. The failure was caused by a delay in OVN performing source network address translation (SNAT) on packets for these instances. Because the IP address for the instance is not translated until OVN pushes the packet, the TCP connection would reset and the link between the instance and the external network was failing. With this update, the bug has been resolved, and the connection between instances and the external network is no longer disrupted.
Story Points: ---
Clone Of: Environment:
Last Closed: 2024-11-21 11:30:57 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: 2262077    
Bug Blocks:    

Comment 4 Ihar Hrachyshka 2024-02-26 15:03:23 UTC
This may be the same scenario as described (and probably fixed) here: https://patchwork.ozlabs.org/project/ovn/patch/20240207155627.806188-1-martin.kalcok@canonical.com/

Comment 12 Lon Hohberger 2024-09-10 17:22:40 UTC
The fix here is to update to a later version of OVN - the rhosp-openvswitch package handles this