AWS. Centos 7.5 m5.large EC2 instance. VPC with 2 subnets. The instance has two network interfaces, one per a subnet. A Network Security Group is configured to allow all TCP inbound and outbound connections from/to my public IP.
[root@ip-10-212-0-116 ~]# cat /etc/iproute2/rt_tables
255 local
254 main
253 default
0 unspec
200 ens5-rt
200 ens6-rt
[root@ip-10-212-0-116 ~]# ip route
default via 10.212.0.1 dev ens5
10.212.0.0/24 dev ens5 proto kernel scope link src 10.212.0.116
10.212.1.0/24 dev ens6 proto kernel scope link src 10.212.1.116
[root@ip-10-212-0-116 ~]# ip add |grep ens
2: ens5: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9001 qdisc mq state UP group default qlen 1000
inet 10.212.0.116/24 brd 10.212.0.255 scope global dynamic ens5
3: ens6: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
inet 10.212.1.116/24 brd 10.212.1.255 scope global ens6
[root@ip-10-212-0-116 ~]# cat /etc/sysconfig/network-scripts/rule-ens{5,6}
from 10.212.0.116/32 table ens5-rt
to 10.212.0.116/32 table ens5-rt
from 10.212.1.116/32 table ens6-rt
to 10.212.1.116/32 table ens6-rt
[root@ip-10-212-0-116 ~]# cat /etc/sysconfig/network-scripts/rule-ens{5,6}
from 10.212.0.116/32 table ens5-rt
to 10.212.0.116/32 table ens5-rt
from 10.212.1.116/32 table ens6-rt
to 10.212.1.116/32 table ens6-rt
So ssh -i Downloads/rasp.pem centos@publicip1
works.
And ssh -i Downloads/rasp.pem centos@publicip2
returns timeout error.
I have done the same configuration in Azure using this guide https://docs.microsoft.com/en-us/azure/virtual-machines/linux/multiple-nics and it works. The difference I have noticed is that in AWS ens5 (primary) interface rewrites BOOTPROTO
setting from static
to dhcp
on boot.