Description of problem: OVS causing high pings and latency inside guest VM when a active DPDK port fails Version-Release number of selected component (if applicable): Red Hat OpenStack Platform release 13.0.7 (Queens) openstack-neutron-openvswitch-12.0.6-3.el7ost.noarch Fri Jun 28 03:59:06 2019 openvswitch-2.9.0-103.el7fdp.x86_64 Fri Jun 28 03:58:54 2019 openvswitch-ovn-central-2.9.0-103.el7fdp.x86_64 Fri Jun 28 03:59:07 2019 openvswitch-ovn-common-2.9.0-103.el7fdp.x86_64 Fri Jun 28 03:58:56 2019 openvswitch-ovn-host-2.9.0-103.el7fdp.x86_64 Fri Jun 28 03:59:06 2019 openvswitch-selinux-extra-policy-1.0-9.el7fdp.noarch Fri Jun 28 03:58:39 2019 python-openvswitch-2.9.0-103.el7fdp.x86_64 Fri Jun 28 03:58:54 2019 How reproducible: Always Steps to Reproduce: 1. Fail an active DPDK port 2. 3. Actual results: Should fail over to other port with out causing any latency Expected results: Additional info:
Created attachment 1672012 [details] grabinfov1.sh to help capturing information while reproducing the issue. Please run the script while running openvswitch-2.9.0-103.el7fdp.bz1809363.3 and provide the ping output files and the generated tarball. Thanks fbl
Created attachment 1672598 [details] grabinfov3.sh to request additional pcap
(In reply to Gowrishankar Muthukrishnan from comment #114) > Created attachment 1672598 [details] > grabinfov3.sh to request additional pcap Please note vhu_ping="vhu19948703-d7" at the top of script if the vhu- interface used for ping is right.
This will be released on FDP 20.D (see the cloned bz#1822198)
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days