Bug 1472854
| Summary: | [fdProd] Update OVS to 2.7.2 | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Timothy Redaelli <tredaelli> |
| Component: | openvswitch | Assignee: | Timothy Redaelli <tredaelli> |
| Status: | CLOSED ERRATA | QA Contact: | ovs-qe |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.4 | CC: | atelang, atragler, ctrautma, fleitner, ovs-team, pezhang, qding |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| URL: | https://mail.openvswitch.org/pipermail/ovs-announce/2017-July/000239.html | ||
| Whiteboard: | |||
| Fixed In Version: | openvswitch-2.7.2-1.git20170719.el7fdp | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2017-08-03 12:35:39 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: | |||
Verified with openvswitch-2.7.2-1.git20170719.el7fdp Reference to https://bugzilla.redhat.com/show_bug.cgi?id=1451911#c10 Live migration with single queue testing: PASS Versions: 3.10.0-693.el7.x86_64 qemu-kvm-rhev-2.9.0-16.el7.x86_64 libvirt-3.2.0-14.el7.x86_64 openvswitch-2.7.2-1.git20170719.el7fdp.x86_64 tuned-profiles-cpu-partitioning-2.8.0-5.el7.noarch dpdk-17.05-3.el7fdb.x86_64 ==Scenario 1: live migration with vhost-user single queue== Check downtime, totaltime, ping/MoonGen packets loss during live migration, ooks good. Host&Guest: 1G hugepage size Guest: 8G memory MoonGen flow: 3Mpps/bidirectional (1) openvswitch acts as vhostuser server mode: No Stream_Rate Downtime(ms) Totaltime(ms) Ping_Loss Moongen_Loss 0 3Mpps 157 12878 18 1121785 1 3Mpps 148 12855 17 1071197 2 3Mpps 156 13035 18 1129271 3 3Mpps 151 12893 16 1093942 4 3Mpps 159 12826 18 1138983 5 3Mpps 161 13065 18 1150769 6 3Mpps 165 12863 19 1179172 7 3Mpps 146 12888 17 1150652 8 3Mpps 165 12869 18 1179632 9 3Mpps 161 12896 18 1163880 (2) openvswitch acts as vhostuser client mode: No Stream_Rate Downtime(ms) Totaltime(ms) Ping_Loss Moongen_Loss 0 3Mpps 154 15419 18 1135606 1 3Mpps 149 15199 16 1088184 2 3Mpps 159 14750 18 1147670 3 3Mpps 162 15580 18 1165776 4 3Mpps 160 15083 17 1154604 5 3Mpps 162 14918 18 1183521 6 3Mpps 155 15485 18 1123589 7 3Mpps 162 15329 18 1153911 8 3Mpps 163 14892 19 1173294 9 3Mpps 161 15719 18 1148970 ==Scenario 2: live migration with vhost-user 2 queues== Will be continued in next Comment.. Live migration with 2 queues testing: PASS ==Scenario 2: live migration with vhost-user 2 queues== Versions: 3.10.0-693.el7.x86_64 qemu-kvm-rhev-2.9.0-16.el7.x86_64 libvirt-3.2.0-14.el7.x86_64 openvswitch-2.7.2-1.git20170719.el7fdp.x86_64 tuned-profiles-cpu-partitioning-2.8.0-5.el7.noarch dpdk-17.05-3.el7fdb.x86_64 Default parameters: Traffic Generator: MoonGen Acceptable Loss: 0.002% Frame Size: 64Byte Bidirectional: Yes Search run time:60s Validation run time: 30s Virtio features: default CPU: Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz NIC: 10-Gigabit X540-AT2 Check throughput value after each migration, looks good. Below is ping-pong migration. (1) openvswitch acts as vhostuser server mode: run 1: 18.519083Mpps (throughput before migration) run 2: 19.084536Mpps (throughput after migrating from host1 to host2 host) run 3: 18.519032Mpps (throughput after migrating from host2 to host1 host) (2) openvswitch acts as vhostuser client mode: run 1: 19.084477Mpps (throughput before migration) run 2: 19.084590Mpps (throughput after migrating from host1 to host2 host) run 3: 19.763328Mpps (throughput after migrating from host2 to host1 host) Note: As qemu bug[1] exists, that's why only check throughput value here. [1]Bug 1450680 - Migrating guest with vhost-user 2 queues and packets flow over dpdk+openvswitch fails: guest hang, and qemu hang or crash 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/RHSA-2017:2418 |
From ovs-announce mailing list: "Due to an issue introduced in the previous release, it is highly recommended that users of 2.7.1 upgrade." commit 9a03615624972142e750484950d7348190d68b8b Author: Han Zhou <zhouhan> Date: Thu Jul 13 23:44:51 2017 -0700 Revert "netdev: Fix netdev_open() to adhere to class type if given" This reverts commit d3b8f5052292b3ba9084ffed097e90b87f2950f5. The commit introduced a problem that "File exists" will be reported when trying to open br0. The operation that adds eth0 to br0 while moving IP address from eth0 to bridge internal interface br0 reproduces this issue. $ ip a del <ip> dev eth0; ip a add <ip> dev br0; ovs-vsctl add-port br0 eth0 $ ovs-dpctl show ... port 1: br0 (internal: open failed (File exists)) ... At this point restarting OVS will result in connection lost for the node. Reverting the change fixes the problem. Since adding physical interface to OVS bridge is quite normal operation, the problem is more severe than the original problem fixed by commit d3b8f5052, so revert this before a better fix is found for the original problem. Signed-off-by: Han Zhou <zhouhan> Signed-off-by: Justin Pettit <jpettit>