-
Bug
-
Resolution: Unresolved
-
Undefined
-
4.16.z
-
Quality / Stability / Reliability
-
False
-
-
None
-
Important
-
None
-
Done
-
Feature
-
-
None
-
None
-
None
-
None
Description of problem:
After OpenShiftSDN to OVN-K live migration with custom NMstate, br0 is left on the system.
- name: br0 type: ovs-interface state: ignore mtu: 1450 ipv4: enabled: false ipv6: enabled: false controller: br0 accept-all-mac-addresses: false
Version-Release number of selected component (if applicable):
4.16.38
How reproducible:
Once
Steps to Reproduce:
1. install OpenShift SDN 4.14.50 and configure balance-slb OVS bonds using nmstate https://github.com/RHsyseng/rhcos-slb/blob/simplify-networking/README.md 2. upgrade to 4.15 3. upgrade to 4.16.38 4. apply OVN-K live migration workarounds https://github.com/rbbratta/rhcos-slb/blob/sdn-to-ovn-migration/sdn-to-ovn/README.md 5. live migrate to OVN-K 6. wait
Actual results:
br0 present
Expected results:
br0 not present
Additional info:
Custom NMstate was used to remove extra brcnv, this occurs after the first reboot?
- name: brcnv type: ovs-interface state: absent
At what point in the live-migration can we remove br0?
- blocks
-
OCPBUGS-59928 after OVN-K live migration br0 is still present
-
- Closed
-
- is cloned by
-
OCPBUGS-59928 after OVN-K live migration br0 is still present
-
- Closed
-
- is duplicated by
-
OCPBUGS-58182 OVN-K live migration fails to node drain due to oauth/apis failures
-
- Closed
-
- is related to
-
OCPBUGS-57625 document OCP reserved bridge names 'br0'
-
- Closed
-
- links to