[SOLVED] Adding non-sequential IP blocks to 'public' range

[TGW]

I have cruised the forums fairly thorough and have not quite found a
solution to my problem. Maybe someone can help . . . Thanks in advance!

I have (2) /29 'non-contiguous' IP blocks to be used as public IPs for
a few containers and 2 websites (VM's).

I have a rough understanding of how the /etc/network/interfaces file works

Can anyone please elaborate on the syntax I would use to accomplish this
task for the above file?

Here is my plan for the IP blocks use:


My Proxmox node is using (.114) IP from the 'A' pool

'A' pool 104.250.xxx.114->118
NM 255.255.255.248
GW 104.250.xxx.113

there are 4 IP's remaining (actually 3) with one being used for the
IPMI login (104.250.xxx.118) and the rest to be used for containers.



Provider gave me another /29 IP block to also be used with this node.

'B' pool 192.200.xxx.34->38
NM 255.255.255.248
GW 192.200.xxx.33

there are a total of 5 IPs to be used here . . .


How would I configure the (2) mentioned /29 blocks to my node so that the
remaining 8 IP's can be used as isolated 'public' IP's for containers and VM's?


Any help with this, is GREATLY appreciated!

 

[TGW]

Unfortunately, above solution does not work. Leaving either the IP or netmask blank results in a nerfed connection . . .
I know there has to be an answer out there . . . Would the IT gods smile upon this IT noob adventurer . . .

 

[bobmc]

So for vmbr2 you can ping .33 (gateway) and .34 (host) on the 192.200.xxx.xxx network - from where (a laptop or desktop? with what IP?). I take it that .35 is an IP assigned to a running VM? Can you ping .35 from the host. Can you ping .34 from the VM?

 

[bobmc]

So the containers and the VM's can ping each other on the 'B' network and they can ping vmbr2 but they can't reach the internet or the 'A' network and they can't ping their own gateway.

Does the internet router have both IP addresses 10.250.x.113 and 192.200.x.33?

If so try adding the 192.200.x.33 gateway to vmbr2 and to each of the CT and VM. Then see if they can ping the gateway and the internet.

 

[bobmc]

OK, So basics here.

1. Do you have two physical connections into your proxmox host
2. Are they plugged into the router or do they go via a switch
3. Are ports unbound at both ends
4. If you have a dhcp scope for physical clients, what is their traceroute to google.com
5. If you connect a physical client to the same hardware as the proxmox host, and give it a 'B' pool IP/gateway, can it reach the internet, what does traceroute show.

 

[TGW]

Bob,

I submitted some of these questions to my host provider, as I do not know the configuration up to the server.

Traceroute shows the following on the 'A' pool routing:

traceroute to www.google.com (172.217.5.196), 30 hops max, 60 byte packets
1 104-250-(xxx)-113.static.gorillaservers.com (104.250.xxx.113) 2.219 ms 2.534 ms 2.817 ms
2 104-250-(xxx-1)-229.static.gorillaservers.com (104.250.xxx-1.229) 0.785 ms 0.894 ms 0.957 ms
3 104-250-(xxx-2)-17.static.gorillaservers.com (104.250.xxx-2.17) 0.776 ms 0.830 ms 0.857 ms
4 199-73-121-241.static.gorillaservers.com (199.73.121.241) 14.682 ms 14.855 ms 14.896 ms
5 199-73-120-237.static.gorillaservers.com (199.73.120.237) 14.583 ms 14.679 ms 14.720 ms
6 100-42-221-1.static.webnx.com (100.42.221.1) 15.939 ms 15.651 ms 15.688 ms
7 google.as15169.any2ix.coresite.com (206.72.210.41) 14.718 ms 14.775 ms 14.740 ms
8 108.170.247.161 (108.170.247.161) 14.565 ms 14.500 ms 14.540 ms
9 216.239.40.169 (216.239.40.169) 14.608 ms 66.249.94.217 (66.249.94.217) 14.598 ms 14.507 ms
10 lax28s10-in-f4.1e100.net (172.217.5.196) 14.488 ms 14.379 ms 14.453 ms

 

[bobmc]

Yeah, that traceroute is as expected, the host traffic goes straight to 104.250.xxx.113 and out to the the internet via 104.250.xxx.229. Everything from that point is external to your system.

What we haven't yet established is whether there is a viable route out via your second 'B' pool gateway 192.200.xxx.33

 

[TGW]

Bob, thank you for your help, I have solved this and I will PM you my solution!

 

[Amit]

TGW, can you post your solution? I have the exact scenario at the moment and am having trouble since I have 2 gateways that I need to use on one host node. Can you help shed some light on how you got this working?