Back to bug 2217867
| Who | When | What | Removed | Added |
|---|---|---|---|---|
| Red Hat One Jira (issues.redhat.com) | 2023-06-27 10:15:59 UTC | Link ID | Red Hat Issue Tracker OSP-26110 | |
| Juan Pablo Marti | 2023-06-27 12:46:23 UTC | CC | hakhande | |
| Jaison Raju | 2023-06-27 15:59:08 UTC | Keywords | Performance | |
| CC | jraju | |||
| Vijayalakshmi Candappa | 2023-06-28 07:16:16 UTC | CC | vcandapp | |
| Gurpreet Singh | 2023-06-28 15:30:54 UTC | CC | gurpsing | |
| Miguel Angel Nieto | 2023-07-20 10:40:08 UTC | CC | mnietoji | |
| Robin Jarry | 2023-07-21 10:46:57 UTC | QA Contact | rhosp-nfv-int | ekuris |
| Flags | needinfo?(mleitner) | |||
| CC | apevec, chrisw, mleitner, rjarry | |||
| Keywords | Triaged | |||
| Component | openstack-nova | openvswitch | ||
| Robin Jarry | 2023-07-21 12:04:55 UTC | Doc Type | If docs needed, set a value | Known Issue |
| Robin Jarry | 2023-07-21 12:14:51 UTC | Doc Text | Cause: Software Managed Flow Steering (smfs) can cause issues with certain types of flows. For example, LLDP offloaded flows can cause performance regression in all VFs associated with the PF where the flow is programmed. Consequence: Performance regression. Workaround (if any): Use Device Managed Flows Steering (dmfs) instead of smfs. Result: Performance is normal. |
|
| Robin Jarry | 2023-07-21 13:09:15 UTC | Comment 2 is private | 1 | 0 |
| Comment 3 is private | 1 | 0 | ||
| Comment 4 is private | 1 | 0 | ||
| Comment 5 is private | 1 | 0 | ||
| Comment 6 is private | 1 | 0 | ||
| Comment 7 is private | 1 | 0 | ||
| Comment 8 is private | 1 | 0 | ||
| Comment 9 is private | 1 | 0 | ||
| Comment 10 is private | 1 | 0 | ||
| Comment 11 is private | 1 | 0 | ||
| Comment 12 is private | 1 | 0 | ||
| Robin Jarry | 2023-07-21 13:12:41 UTC | Group | nvidia | |
| Robin Jarry | 2023-07-21 13:23:07 UTC | Group | nvidia, redhat, private | |
| Robin Jarry | 2023-07-21 13:42:36 UTC | CC | atzin, mhazan, mkabat | |
| Group | mellanox, redhat, nvidia | |||
| Marcelo Ricardo Leitner | 2023-07-21 14:19:42 UTC | Blocks | 2172622 | |
| Marcelo Ricardo Leitner | 2023-07-21 20:12:37 UTC | Flags | needinfo?(mleitner) | needinfo?(rjarry) |
| Robin Jarry | 2023-07-25 13:18:12 UTC | Flags | needinfo?(rjarry) | |
| Robin Jarry | 2023-07-25 13:20:17 UTC | Summary | [OSP-17.1][Mellnox-Cx6] OVN-HWOL Performance is lower compared to Regular SR-IOV Performance | [OSP-17.1][Mellnox-Cx6] OVN-HWOL: LLDP flows cause performance regression |
| Robin Jarry | 2023-07-27 12:39:40 UTC | Group | nvidia, mellanox, redhat | |
| Robin Jarry | 2023-08-02 11:48:45 UTC | Doc Text | Cause: Software Managed Flow Steering (smfs) can cause issues with certain types of flows. For example, LLDP offloaded flows can cause performance regression in all VFs associated with the PF where the flow is programmed. Consequence: Performance regression. Workaround (if any): Use Device Managed Flows Steering (dmfs) instead of smfs. Result: Performance is normal. | Cause: There is a know issue on Nvidia ConnectX-5 and ConnectX-6 NICs when using hardware offload. Some offloaded flows on a PF can cause transient performance regressions on its associated VFs. This specific regression was observed with LLDP and VRRP traffic. Consequence: Performance regression. Workaround (if any): None. Result: |
| Robin Jarry | 2023-08-02 11:54:28 UTC | Flags | needinfo?(bnemeth) needinfo?(wizhao) | |
| CC | bnemeth, wizhao | |||
| Paul Grist | 2023-08-02 12:38:12 UTC | Status | NEW | ASSIGNED |
| CC | pgrist | |||
| William Zhao | 2023-08-02 16:21:41 UTC | Flags | needinfo?(wizhao) | |
| Robin Jarry | 2023-08-03 09:23:47 UTC | Target Release | --- | 17.1 |
| Target Milestone | --- | z3 | ||
| Ian Frangs | 2023-08-03 15:46:23 UTC | Flags | needinfo?(rhosp-nfv-int) | |
| CC | rhosp-nfv-int | |||
| Sanjay Upadhyay | 2023-08-07 19:37:30 UTC | CC | supadhya | |
| Erin Peterson | 2023-08-09 18:16:12 UTC | CC | erpeters | |
| Doc Text | Cause: There is a know issue on Nvidia ConnectX-5 and ConnectX-6 NICs when using hardware offload. Some offloaded flows on a PF can cause transient performance regressions on its associated VFs. This specific regression was observed with LLDP and VRRP traffic. Consequence: Performance regression. Workaround (if any): None. Result: | There is currently a known issue on Nvidia ConnectX-5 and ConnectX-6 NICs, when using hardware offload, where some offloaded flows on a PF can cause transient performance issues on the associated VFs. This issue is specifically observed with LLDP and VRRP traffic. |
Back to bug 2217867