-
Bug
-
Resolution: Done
-
Undefined
-
None
-
4.12.0
-
None
-
Quality / Stability / Reliability
-
False
-
-
None
-
Important
-
None
-
None
-
Rejected
-
SDN Sprint 225, SDN Sprint 226
-
2
-
None
-
Bug Fix
-
NA
-
None
-
None
-
None
-
None
Description of problem:
Opening this BZ to track this issue. Following rule was created and xdp program was attached to the interface
ens192: <BROADCAST,MULTICAST,ALLMULTI,UP,LOWER_UP> mtu 1500 xdpgeneric qdisc mq master ovs-system state UP mode DEFAULT group default qlen 1000
link/ether 00:50:56:ac:ca:fb brd ff:ff:ff:ff:ff:ff
prog/xdp id 2 tag ef65c8e7d746da72 jited
Rule
--------
ingress:
- rules:
- action: Deny
order: 1
protocolConfig:
protocol: TCP
tcp:
ports: 30321-33000
sourceCIDRs:
- 10.0.5.26/12
interfaces:
- ens192
nodeSelector:
matchLabels:
node-role.kubernetes.io/worker: ""
status:
syncStatus: Synchronized
kind: List
Version-Release number of selected component (if applicable):
4.12
How reproducible:
rarely
Steps to Reproduce:
1. A rule was created to Deny traffic at a particular TCP port 2. daemons ds and controller manager deployment was deleted. Hence they re-spawned sucessfully 3. the same rule at step 1 was applied but with Allow rule 4. step 2 repeated 5. Deny rule was reapplied but its allowing the traffic at that port
Actual results:
rule becomes non functional post daemons ds and controller manager deletion and they re spawned successfully
Expected results:
Post manager and daemons redeployment, rule should become functional again
Additional info: