Vm not ping after migration on cluster's three nodes

[Dyan]

Hi everybody,

I've make a cluster of proxmox PVE 7.x.

My devices :
- PVE1 : server Xeon with 32gb ram and 2x1 TB with raid mdadm. + 1 ipfailover.
- PVE2 : server Atom octo with 1x240 SDD and 16gb ram.
-PVE3 : server Atom octo with 1x240 SDD and 16gb ram in a datacenter 1000km away.
- RPN (Similar Vrack at ovh)


My work :
I resized mdadm raid partitions to create ZFS storage. Idem for two little servers.
I have configured my virtual private network (RPN) the ping machines and do multicast. I created my cluster.
I created on my PVE1 machine a deb11 LXC container to test, I gave it my ipfailover + MAC address. Everything works it ping from the outside great!
But when I migrate my LXC container from PVE1 to PVE2 or PVE3, the migration result is correct, but my container does not ping from outside.


Do you have an idea of the problem ? Thank you soo much for your help.

Have a nice day.

 

[supermix]

Hi friend prolly ip failover assign to PVE1. U need assign failover to destination Node.

 

[Dyan]

Hi Friend,

Thank for your reply.
What do you mean ?

Precision, the ip failover is attached to the PVE1 server on my dashboard of my provider

This morning i have tested this configuration :
For PVE1 networking :
eno1 this primary network card
eno2 this secondary network card for private network of my provider.
vmbr0 this Linux Bridge of eno1 for my public ip : 130.130.130.130/24 and : 130.130.130.1/24 gateway.

For LXC container :
name : eth0
bridge : vmbr0
Mac Address : The one provided by provider.
Ip address : IPFO : 150.150.150.150/24
Passerelle : IPFO : 150.150.150.1

The LXC container work on PVE1 after migration to PVE2 it's not working.

For PVE2 networking :
enp0s20f0 this primary network card
enp0s20f1 this secondary network card for private network of my provider.
vmbr0 this Linux Bridge of eno1 for my public ip : 164.164.164.164/24 and gateway :164.164.164.1.

Thank you for your help, see you later.

 

[supermix]

hi friend i am see your failover hace own gateway... in this case i assume u are using Public IP block linked to OVH Vrack??

 

[Dyan]

Hi Supermix,

Not even, I added .1 at the end of my ipfo. If I replace my gateway on the LXC container with the gateway from my PVE1 or PVE2, it pings when the LXC is on PVE1.

 

[supermix]

yes ok u must setup right gateway on LXC container depending where is running at any time

 

[Dyan]

I put the PVE2 gateway on the container after migration + reboot. it does not ping.

 

[supermix]

I put the PVE2 gateway on the container after migration + reboot. it does not ping.

u need check where is assigned failover on your provider dashboard

 

[Dyan]

I just checked It is assigned on PVE1 and it has the same gateway as PVE1. I can't move the IP failover to another machine from my provider's dashboard. My ipfailover is locked on server.