-
Bug
-
Resolution: Unresolved
-
Major
-
None
-
4.18.z
-
Quality / Stability / Reliability
-
False
-
-
None
-
Critical
-
Yes
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
Install cluster upgrade to 4.18.26 Create first must-gather and natlists delete EgressIPs, rebuild DB, redeploy EgressIPs Create second must-gather and natlists Find pod with networking issues and create inspect from its namespace.
Version-Release number of selected component (if applicable):
4.18.26 (upgraded and clean installations)
How reproducible:
multiple clusters replicated on customer environment
Steps to Reproduce:
Install cluster upgrade to 4.18.26
Create first must-gather and natlists
delete EgressIPs, rebuild DB, redeploy EgressIPs
Create second must-gather and natlists
Find pod with networking issues and create inspect from its namespace.
[see first comment below for extended replication details + highlights]
Actual results:
egressIP traffic is unreliable/unavailable even after DB rebuild + mitigation steps provided from previous condition failure in 4.18.23
Expected results:
egressIP reliability
Additional info:
This issue is likely a regression - we have updated from 4.18.22 --> 4.18.23 which alleviated problem conditions for this platform. Then Upgraded to 4.18.26 and observed reintroduction of problem condition. (subsequently replicated again on a separate new cluster installation as well).
