VMs unable to ping each other

[uglybotak18]

Dear Team,

I have tried to setup 2 servers, and create VM in each server
VM1 is not able to ping VM2 and vice versa VM2 is not able to ping VM1
they have no issues in accessing internet

eth0 - 10.10.10.35/29
vmbr0 - 68.231.45.35/29 IPv4/CIDR
gateway : 68.231.45.33


eth0 - 10.10.10.139/29
vmbr0 : 68.231.45.139/29 IPv4/CIDR
gateway : 68.231.45.137

any advise to rectify?

Thank you

 

[Lorenz.S]

The VMs and gateways are in different networks. Do these gateways (I assume the servers) have a route to each other?
Try ip route on each

 

[Dominic]

Hi,

you can additionally use tcpdump on different interfaces to see where packages arrive. For example,

tcpdump icmp -i vmbr0

 

[uglybotak18]

The VMs and gateways are in different networks. Do these gateways (I assume the servers) have a route to each other?
Try ip route on each

Hi Lorenz,

Where do I find the reference for this routing ?

Thank you

 

[Lorenz.S]

Please post your network configuration (/etc/network/interfaces) and the output of the ip route command, so we may understand your setup better. Reference for ip route: https://www.systutorials.com/docs/linux/man/8-ip-route/

 

[uglybotak18]

root@firstnode:/etc/network# cat /etc/network/interfaces
auto lo
iface lo inet loopback

iface enp5s0f0 inet manual

auto enp5s0f1
iface enp5s0f1 inet static
address 10.10.10.35/29

auto vmbr0
iface vmbr0 inet static
address 68.231.45.35/29
gateway 68.231.45.33
bridge-ports enp5s0f0
bridge-stp off
bridge-fd 0


root@firstnode:/etc/network# ip route
default via 68.231.45.33 dev vmbr0 onlink
10.10.10.32/29 dev enp5s0f1 proto kernel scope link src 10.10.10.35
68.231.45.32/29 dev vmbr0 proto kernel scope link src 68.231.45.35

=====

root@secondnode:~# cat /etc/network/interfaces
auto lo
iface lo inet loopback

iface enp6s0f0 inet manual

auto enp6s0f1
iface enp6s0f1 inet static
address 10.10.10.139/29

auto vmbr0
iface vmbr0 inet static
address 68.231.45.139/29
gateway 68.231.45.137
bridge-ports enp6s0f0
bridge-stp off
bridge-fd 0


root@secondnode:~# ip route
default via 68.231.45.137 dev vmbr0 onlink
10.10.10.136/29 dev enp6s0f1 proto kernel scope link src 10.10.10.139
68.231.45.136/29 dev vmbr0 proto kernel scope link src 68.231.45.139

 

[Lorenz.S]

Is the gateway one gateway with two interfaces, or two separate gateways? If the latter, do they have a route to each other?

Another question: what is the intended purpose of enp6s0f1/enp5s0f1?

 

[uglybotak18]

Hi Lorenz,

is 2 separate gateway. We getting 2 bare metals from vendor, which they provide 2 different set of ip address ( 2 gateway)
How do I route the 2 different gateway? any reference link for me to look up?

enp6s0f1/enp5s0f1 is the network interface card.
I suppose to have 1 public network where VM are able to ping each other,
1 private network for ceph to work ( where im still implementing
How do i set in the public network which allow VM to ping each other?

Many thanks in advance

 

[Lorenz.S]

Thank you for your reply

Before any further action, I would suggest to test if

• VM on hypervisor "A" (VMa) can ping gateway for hypervisor "A" (GWa) and GWb.
• VMb can ping GWb and GWa.
• GWa can ping GWb.

If the last one is not working, it is probably a routing problem and not PVE/Proxmox related. (see routing or static routing on Wikipedia; configuration depends on vendor and OS of gateways – Google is your friend)
If the last one is working, we have to look for other problems.