https://rancher.com/ logo
#longhorn-storage
Title
# longhorn-storage
a

adamant-kite-43734

03/08/2022, 11:11 PM
This message was deleted.
b

bland-painting-61617

10/02/2022, 9:00 AM
Wanted to bring this topic back up,
Allow Node Drain with the Last Healthy Replica
is on but kured is unable to drain the node - there is one volume left on the node but it should be rebooted. Yesterday another node was stuck in this state with no volumes on the node itself - a bug?
Copy code
evicting pod longhorn-system/instance-manager-e-9ca60819
error when evicting pods/"instance-manager-e-9ca60819" -n "longhorn-system" (will retry after 5s): Cannot evict pod as it would violate the pod's disruption budget.
error when evicting pods/"instance-manager-r-850386c0" -n "longhorn-system" (will retry after 5s): Cannot evict pod as it would violate the pod's disruption budget.
b

billowy-painting-56466

10/03/2022, 1:28 AM
When a node is marked unschedulable, Longhorn checks if the PBD is deletable before deleting the PDB. From the thread description it seems all the conditions are met and somehow PDB is not deleted. 1. nodes are marked unschedulable 2. no volumes are attached to that node. 3. Then setting Allow Node Drain with the Last Healthy Replica set to true. I suggest you to create an issue with the reproducing steps and support bundle. Ref to similar issue, https://github.com/longhorn/longhorn/issues/2639.
b

bland-painting-61617

10/03/2022, 9:12 AM
Yup, thanks, I raised the issue as I thought it would be a more accountable way for the team to manage.
👍 1
347 Views