Hi,
I've a production cluster with 5 nodes and in a random fashion I've had some NICs stop receiving traffic outta nowhere.
Last example today, via a tcpdump i see that the NIC correctly passes packets from VM to outside, while it doesn't receive any traffic. I can easily solve the issue by re-generate the NIC mac address, and after I've done it, i can reuse that mac address on the same VM or another one while if that's still stuck, i can live migrate the VM from one node to another, and it remains stuck.
Network configuration is:
Bond from switch to node, bridge on that bond, and interfaces on that bridge
Have you ever faced a problem like this? Any help?
Versions:
proxmox: 6.4-13
pve-qemu-kvm_5.2.0
ceph 15.2.17
I know those are outdated, but if there's a bug i need to provide evidence to my manager to upgrade the cluster with proper reasons and a changelog which fixes that bug (will it be debian, pve, kvm or whatever)
Thanks in advance!
I've a production cluster with 5 nodes and in a random fashion I've had some NICs stop receiving traffic outta nowhere.
Last example today, via a tcpdump i see that the NIC correctly passes packets from VM to outside, while it doesn't receive any traffic. I can easily solve the issue by re-generate the NIC mac address, and after I've done it, i can reuse that mac address on the same VM or another one while if that's still stuck, i can live migrate the VM from one node to another, and it remains stuck.
Network configuration is:
Bond from switch to node, bridge on that bond, and interfaces on that bridge
Have you ever faced a problem like this? Any help?
Versions:
proxmox: 6.4-13
pve-qemu-kvm_5.2.0
ceph 15.2.17
I know those are outdated, but if there's a bug i need to provide evidence to my manager to upgrade the cluster with proper reasons and a changelog which fixes that bug (will it be debian, pve, kvm or whatever)
Thanks in advance!