[BUG] Test V2 Data Engine Selective Activation With Existing Engine And Replica failed #12539
Description
Describe the Bug
Test Test V2 Data Engine Selective Activation With Existing Engine And Replica failed on v1.10.2-rc1. This case passed on v1.11.x and can be manually reproduced
Labeling a node that already has a v2 volume engine attached with command kubectl label node ${node_0} node.longhorn.io/disable-v2-data-engine=true should be rejected, but on v1.10.2-rc1 it succeeds.
# k get nodes
NAME STATUS ROLES AGE VERSION
ip-10-0-1-167 Ready control-plane,master 38m v1.33.4+k3s1
ip-10-0-2-11 Ready <none> 35m v1.33.4+k3s1
ip-10-0-2-174 Ready <none> 35m v1.33.4+k3s1
ip-10-0-2-206 Ready <none> 35m v1.33.4+k3s1
#
# k get node ip-10-0-2-11 -o yaml | grep node.longhorn.io/disable-v2-data-engine
node.longhorn.io/disable-v2-data-engine: "true"
#
# kl get engine
NAME DATA ENGINE STATE NODE INSTANCEMANAGER IMAGE AGE
vol1-e-0 v2 running ip-10-0-2-11 instance-manager-03d67f246d87799a5d034c624453147f docker.io/longhornio/longhorn-instance-manager:v1.10.x-head 14m
To Reproduce
Run
-t \"Test V2 Data Engine Selective Activation With Existing Engine And Replica\" --exclude \"cluster\" --exclude \"storage-network\" --exclude \"large-size\" -v LOOP_COUNT:1 -v RETRY_COUNT:1200 -v DATA_ENGINE:v2
Or
- Attach a v2 volume to node 0
- Wait the volume healthy
- Run command
kubectl label node <node 0> node.longhorn.io/disable-v2-data-engine=true - Command success but should not
Expected Behavior
kubectl label node <node 0> node.longhorn.io/disable-v2-data-engine=true should fail with message like cannot disable v2 data engine
Support Bundle for Troubleshooting
supportbundle_da8e11a0-ea21-41c6-a564-5b500ee1a29b_2026-01-26T01-03-40Z.zip
Environment
- Longhorn version: v1.10.x
- Impacted volume (PV): vol1
Additional context
Might related to #11831
Workaround and Mitigation
No response
Metadata
Metadata
Assignees
Labels
Type
Projects
Status