please delete this post

[kemil]

Hi there,

I installed ifupdown2 and now server cannot restart in HD mode, It is in rescue mode and all network configuration is gone.

interfaces and interface.d/ is empty.

How can I revert back the network configuration or please instruct on how to create a configuration file.

Thanks.

 

[Dunuin]

you can run ip addr to find out your NIC names. Here you will find some examples how a /etc/network/interfaces could look like: https://pve.proxmox.com/wiki/Network_Configuration#_default_configuration_using_a_bridge
You can try new network configs without rebooting with systemctl restart networking.

 

[kemil]

Could you please help me find the right configuration? I do not have DHCP, but for other dedicated server if I go to the interfaces file then I see below written configuration, so I tried it like that.

as you can see below no network bridge is connected to any nic. I am new to this networking configuration, so if you can please elaborate which is bridge and which is nic. and how should my config look?

/etc/network/interfaces

# network interfaces
auto lo
iface lo inet loopback

iface bond0 inet manual

auto eth0
iface eth0 inet dhcp
bridge-ports bond0
bridge-stp off
bridge-fd 0


but server does not reboot. cannot ping nothing. please help.

 

[Dunuin]

/etc/network/interfaces

# network interfaces
auto lo
iface lo inet loopback

iface bond0 inet manual

auto eth0
iface eth0 inet dhcp
bridge-ports bond0
bridge-stp off
bridge-fd 0

That doesn't make much sense. That bond insn't bonging NICs. And that bridge option should belong to a bridge, not a NIC.

but for other dedicated server if I go to the interfaces file then I see below written configuration, so I tried it like that.

What does the /etc/network/interfaces of your other PVE nodes look like? Were they similar configures?

o if you can please elaborate which is bridge and which is nic.

Default bridge is vmbr0. NIC depends on your server. I guess the NICs eth0 or eth1 were used for management and not the two infiniband NICs.

 

[kemil]

That doesn't make much sense. That bond insn't bonging NICs. And that bridge option should belong to a bridge, not a NIC.

I cannot figure out which one is bridge and which one is nic? what should I write in place of "eth0" and "bond0" ??

What does the /etc/network/interfaces of your other PVE nodes look like? Were they similar configures?

Yes, it is exactly similar to the one I wrote.

Default bridge is vmbr0. NIC depends on your server. I guess the NICs eth0 or eth1 were used for management and not the two infiniband NICs.

so should I write vmbr0 instead of eth0? and i did not understand the later part of this about NICs?

 

[Dunuin]

es, it is exactly similar to the one I wrote.

You mean that one?

/etc/network/interfaces

# network interfaces
auto lo
iface lo inet loopback

iface bond0 inet manual

auto eth0
iface eth0 inet dhcp
bridge-ports bond0
bridge-stp off
bridge-fd 0

Because that one doesn't make sense. What is cat /etc/network/interfaces of the working server returning?

 

[alexskysilk]

Allow me to help

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports eth0
    bridge-stp off
    bridge-fd 0

 

[Dunuin]

Allow me to help

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports eth0
    bridge-stp off
    bridge-fd 0

Maybe also something like this in case a bond/vlan were used:

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

iface eth1 inet manual

iface bond0 inet manual
    bond-slaves eth0 eth1
    bond-miimon 100
    bond-mode 802.3ad
    bond-xmit-hash-policy layer2+3

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports bond0
    bridge-stp off
    bridge-fd 0
    bridge-vlan-aware yes
    bridge-vids 2-4094

 

[kemil]

You mean that one?

Yes this one.
--------
/etc/network/interfaces

# network interfaces
auto lo
iface lo inet loopback

iface bond0 inet manual

auto eth0
iface eth0 inet dhcp
bridge-ports bond0
bridge-stp off
bridge-fd 0

Because that one doesn't make sense. What is cat /etc/network/interfaces of the working server returning?

this is working server details:

/etc/network/interfaces

# network interfaces
auto lo
iface lo inet loopback

iface eno1 inet manual

auto vmbr0
iface vmbr0 inet dhcp
bridge-ports eno1
bridge-stp off
bridge-fd 0

 

[kemil]

Allow me to help

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports eth0
    bridge-stp off
    bridge-fd 0

I am trying this and then restarted network service with,
systemctl restart networking.service

now, I am rebooting it with HD mode, let's see if this helps.

This did not help, the server still has error rebooting it.

 

[kemil]

Maybe also something like this in case a bond/vlan were used:

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

iface eth1 inet manual

iface bond0 inet manual
    bond-slaves eth0 eth1
    bond-miimon 100
    bond-mode 802.3ad
    bond-xmit-hash-policy layer2+3

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports bond0
    bridge-stp off
    bridge-fd 0
    bridge-vlan-aware yes
    bridge-vids 2-4094

should I just try this exactly like this??

 

[kemil]

Allow me to help

# network interfaces
auto lo
iface lo inet loopback

iface eth0 inet manual

auto vmbr0
iface vmbr0 inet dhcp
    bridge-ports eth0
    bridge-stp off
    bridge-fd 0

Hey, I tried this but did not work. If you could check and let me know what could be wrong.

 

[Dunuin]

You could try it with eth1 in case the enternet cable is plugged in eth1 and not in eth0.

Hard to help you if you don't give us more information about your network infrastructure.

 

[kemil]

You could try it with eth1 in case the enternet cable is plugged in eth1 and not in eth0.

Hard to help you if you don't give us more information about your network infrastructure.

because I do not have access to the GUI, I do not know how can I give you info about network infra. all network component names are changed after hard reboot. So, now I cannot figure out looking at other working servers too. The configuration was exactly same as I shared working server details.

there was no vmbr0. please note, I was just trying different things in /etc/network/interfaces to fix and that is when I created this thread, so in the post you could see vmbr0 but it was actually not there when server got into rescue mode.

So, from the fresh start. please see this. after restart again in rescue mode,

now, if you could suggest anything??? I could give you more information if you let me know what and how can I get them for you?

 

[vesalius]

You say servers are exactly the same, but the server with the network down list ifb0 and ifb1 as available interfaces. Those are not listed on the working server with ip addr. Did you recently add a PCI card to this non-working server? Once you get the machine working again you really should switch to a static IP address instead of DHCP.

in the past on one of my nodes the upgrade/change to ifupdown2 from ifupdown left the networking service no longer enabled at boot. May not be the case with you, but sudo systemctl is-enabled networking should return enabled. If not, sudo systemctl enable networking fixed it for me.

 

[Dunuin]

I don't only mean the server itself. You could have setup the switch to work with bonds or VLANs. You got multiple NICs but didn`t told us what NIC is connected to which device and so on. Especially with bonds and VLANs there could be millions of possible combinations.

 

[kemil]

You say servers are exactly the same, but the server with the network down list ifb0 and ifb1 as available interfaces. Those are not listed on the working server with ip addr.

Yes, that is why I am confused, It was all same with all my servers when I had access to this not working servers GUI. it was same as the working server
vmbr0 as bridge and eno1 and eno2 as network devices. but after reboot I could not understand what is what.

Did you recently add a PCI card to this non-working server? Once you get the machine working again you really should switch to a static IP address instead of DHCP.

I actually work remotely, so do not know everything about it. It was set up by someone else. It was working fine. but I installed ifupdown2 and then it messed up.

in the past on one of my nodes the upgrade/change to ifupdown2 from ifupdown left the networking service no longer enabled at boot. May not be the case with you, but sudo systemctl is-enabled networking should return enabled. If not, sudo systemctl enable networking fixed it for me.

Yes, it returns enabled.

 

[kemil]

I don't only mean the server itself. You could have setup the switch to work with bonds or VLANs. You got multiple NICs but didn`t told us what NIC is connected to which device and so on. Especially with bonds and VLANs there could be millions of possible combinations.

I do not have physical access to server, working remotely.

 

[vesalius]

Well looks like the non-working server is now different than the working version. Does someone else have access and passwords to your server? An update to ifupdown2 will not add a bond0 or ifb1 and ifb2 interfaces.

 

[kemil]

Well looks like the non-working server is now different than the working version. Does someone else have access and passwords to your server? An update to ifupdown2 will not add a bond0 or ifb1 and ifb2 interfaces.

No, no one has access at the moment. There is no way I can revert back to how it was before?