Hey all,
so I just setup my two node cluster and spun up a pfSense VM.
The pfSense VM has two virtual network adapters, one is untagged, the other one has a vlan 100 tag.
The cluster is working fine, until I start the pfSense VM. As soon as the vm uses the nic with the vlan tag 100, the other pve node turns red and states that it is not connected.
If I don't use any VLANs, the cluster keeps working, so I already was alble to narrow it down to the VLAN.
Of course I set the Flag, that the Host Network is Vlan aware.
Any idea how I can trace the problem further or even better, solve it altogether?
EDIT:
I can ping -f the other node without any problems, after starting the VM.
Are there any services which I should check and maybe restart?
I am using default Linux Bridges, not OVS.
EDIT2:
Alright, I did some further investigation and found out, that while both nodes are doing a tcp dump, the cluster is back up! So it seems, as soon as both interfaces are in promicious mode, the cluster starts working again.
Multicast seems to be working:
But I am still looking for a fix..
EDIT3:
Maybe it was indeed a Multicast Issue. I enabled "IGMP Querying" on my switch and for now the cluster is working. I will keep a look on it, but if anyone has had a similar issue, please let me know how you fixed it.
Thanks
so I just setup my two node cluster and spun up a pfSense VM.
The pfSense VM has two virtual network adapters, one is untagged, the other one has a vlan 100 tag.
The cluster is working fine, until I start the pfSense VM. As soon as the vm uses the nic with the vlan tag 100, the other pve node turns red and states that it is not connected.
If I don't use any VLANs, the cluster keeps working, so I already was alble to narrow it down to the VLAN.
Of course I set the Flag, that the Host Network is Vlan aware.
Any idea how I can trace the problem further or even better, solve it altogether?
EDIT:
I can ping -f the other node without any problems, after starting the VM.
Are there any services which I should check and maybe restart?
I am using default Linux Bridges, not OVS.
EDIT2:
Alright, I did some further investigation and found out, that while both nodes are doing a tcp dump, the cluster is back up! So it seems, as soon as both interfaces are in promicious mode, the cluster starts working again.
Multicast seems to be working:
Code:
omping -c 10000 -i 0.001 -F -q pve0 pve1
Code:
pve1 : unicast, xmt/rcv/%loss = 9464/9464/0%, min/avg/max/std-dev = 0.055/0.122/0.249/0.025
pve1 : multicast, xmt/rcv/%loss = 9464/9464/0%, min/avg/max/std-dev = 0.058/0.134/0.261/0.027But I am still looking for a fix..
EDIT3:
Maybe it was indeed a Multicast Issue. I enabled "IGMP Querying" on my switch and for now the cluster is working. I will keep a look on it, but if anyone has had a similar issue, please let me know how you fixed it.
Thanks
Last edited: