Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 2017803

Summary: br-int doesn't pass DHCP traffic from VM to br-tun
Product: Red Hat OpenStack Reporter: Alex Stupnikov <astupnik>
Component: openstack-neutronAssignee: Slawek Kaplonski <skaplons>
Status: CLOSED DUPLICATE QA Contact: Eran Kuris <ekuris>
Severity: high Docs Contact:
Priority: high    
Version: 16.1 (Train)CC: aenrique, alisci, bcafarel, ccamposr, chrisw, jlibosva, madgupta, ralonsoh, scohen, skaplons, srengan
Target Milestone: ---Keywords: Triaged
Target Release: ---   
Hardware: All   
OS: All   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2022-01-03 19:38:24 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 Alex Stupnikov 2021-10-27 12:55:31 UTC
Description of problem:

One of VM's in customer's RHOSP 16.1 ML2/OVS w DVR deployment can't obtain DHCP address after its lease expired: there are no answers from Neutron DHCP servers for inbound instance's DHCP Requests.

We have tried to isolate the problem during remote session by running "dhclient eth0" command inside affected VM and collecting traffic dumps. Here are our observations:

1. DHCP requests are seen on inbound tap interface (tcpdump is used)
2. DHCP requests are seen on appropriate linux bridge which has appropriate tap interface as a slave (tcpdump is used)
3. DHCP requests are seen on inbound br-int interface (ovs-tcpdump is used)
4. DHCP requests are not seen on outbound br-int interface (patch-tun, ovs-tcpdump is used)

I will provide UUIDs privately, sosreport from affected compute is attached to support case.


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

RHOSP 16.1.6

How reproducible:

Unclear: problem is sporadic, but we have affected instance available for troubleshooting right now