Bug 1954659
| Summary: | With OVN SB DB being busy, not all claims by ovn-controller are written to the DB | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux Fast Datapath | Reporter: | Jakub Libosvar <jlibosva> |
| Component: | ovn2.13 | Assignee: | OVN Team <ovnteam> |
| Status: | NEW --- | QA Contact: | Jianlin Shi <jishi> |
| Severity: | high | Docs Contact: | |
| Priority: | medium | ||
| Version: | FDP 21.B | CC: | ctrautma, eduen, ffernand, jishi, m.andre, mmichels, ralongi, rkhan |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 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 of problem: We have a flaky environment with VIP failing over in between 3 VMs on different chassis. Here are the events how they go in time: The node 1 claims the vip: 2021-04-28T11:25:56.885Z|05691|pinctrl|INFO|Claiming virtual lport 560d0404-d880-436c-85a1-6d837a5ee656 for this chassis with the virtual parent 20e3c9f2-2146-48ec-aa64-670ed3e89878 SB DB is updated record 33173: 2021-04-28 11:26:14.693 table Port_Binding row 6cc4860e (6cc4860e): chassis=24e23da4-822b-49ff-b130-46825b732baf virtual_parent="20e3c9f2-2146-48ec-aa64-670ed3e89878" The node 2 claims the vip: 2021-04-28T11:27:57.488Z|03372|pinctrl|INFO|Claiming virtual lport 560d0404-d880-436c-85a1-6d837a5ee656 for this chassis with the virtual parent 73983cf1-3d3b-4660-ad15-95baffd9a4d3 The node 3 claims the vip and it stays there: 2021-04-28T11:28:31.156Z|09316|pinctrl|INFO|Claiming virtual lport 560d0404-d880-436c-85a1-6d837a5ee656 for this chassis with the virtual parent 22ccfce2-5170-40aa-a29b-6ec0654edf4c SB DB is updated to the node 2 record 33181: 2021-04-28 11:28:53.843 table Port_Binding row 6cc4860e (6cc4860e): chassis=707f372a-5e47-41bb-9c80-fa25c609e739 virtual_parent="73983cf1-3d3b-4660-ad15-95baffd9a4d3" There is no other update to the port binding and it stays like this. From this point the traffic is delivered to the node 2 despite the VIP being on node 3. Version-Release number of selected component (if applicable): ovn2.13-20.12.0-104.el8fdp.x86_64 How reproducible: Not sure Steps to Reproduce: 1. Have a busy SB DB 2. Migrate VIP around 3. Observe the parent port of the VIP and if it's on a correct node Actual results: Expected results: Additional info: NB and SB DBs are attached