This seems very general and would be simple, but I am not able to realize
I have few public ip from isp a.b.c.d/29, the first of range is as gateway, the second assigned to my rb’s first ethernet as uplink.
Other ethers are in bridge where stays a pppoe server, where all clients are connected with a private subnet 10.1.1.2 and go on, these ip are masqueraded, and this schema is working.
Now, I want to give the third public ip address to a particular client, but no success. I put the ip in the client ppp secret setup as remote , it can connect and the ip is really assigned to it, but no internet. It can ping server 10.1.1.1 and the public ip of uplink, but nothing else
It do not seem to me the case of ospf, because I own 2-3 public ip only, which are continuous in the same subnet of gw-uplink, not like the sample of ospf, which uses an entire /24 range, completely different from gw/uplink range
The arp-proxy method is too invasive, I think.
Make one schema with all device on your network,from your isp gateway to your pppoe-server, to understand better your network, and write real number except first two, like a.b.123.44/29
of course, range is: a.b.37.216/29. the ip in the rb ether1-uplink is: a.b.37.218/29, gateway is: a.b.37.217.
(ip route print
DST-ADDRESS PREF-SRC GATEWAY DISTANCE
0 A S 0.0.0.0/0 a.b.37.217 1
)
all other ethernets are in bridge1 10.1.1.1. There is ip pool: 10.1.1.2-10.1.1.254, a pppoe server on bridge1 with this pool, and several pppoe secrets(clients)
10.1.1.0/24 is masqueraded.
Now, I want to assign the a.b.37.219 to a specific pppoe client. I put this ip to the “remote address” in ppp secret properties, it has this address when connected, but it is not connected to internet, as I previously.
you must catch all Ip address on the routerboard (assign all available ip to ether1) and use NAT for netmask on 10.1.1.x
remember on route 0.0.0.0/24 → 66.6.26.217 to set pref-src to 66.6.26.218
like (for each addresses)
/ip firewall nat
add action=netmap chain=dstnat dst-address=66.6.26.219 to-addresses=10.1.1.219
add action=netmap chain=srcnat src-address=10.1.1.219 to-addresses=66.6.26.219
OR
if you have access to gateway 66.6.26.217, add the route
66.6.26.219 by 66.6.26.218
66.6.26.220 by 66.6.26.218
66.6.26.221 by 66.6.26.218
66.6.26.222 by 66.6.26.218
and add public ip to the ppp client profile
OR
use EoIP with one end on bridge with ether1 on gateway and the other end as internet gateway for ppp client.
The second method I have described is the better on absolute.
About EoIP:
Create one bridge on routerboard server and put inside ether1 and one EoIP for each public IP available.
(and assign IP 66.6.26.218 to bridge, instead of ether1)
Create EoIP end on client you want assign the public IP,
assign the public IP on EoIP end on client, like 66.6.26.219
remove default route added automatically by ppp and add route 0.0.0.0/0 to public ip address 66.6.26.217
now the client to go on internet pass on EoIP that are passing on ppp tunnel, respecting bandwidth limit of ppp tunnel, and are on “switch” with gateway 217 and routerboard 218.
It’s like physical cable connected from client to your isp gateway.
There are also other metod someone can suggest, really are more than only one way to do the same thing on RouterOS…
Remember to add Karma to me, if I solved your problem! Thanks.
unfortunately, I haven-t access at gw router, it is of isp, I may ask to whom but let see
I am trying to realize this, do not understand this sentence, it is on bridge with ether1, it has local address and remote address, pppoe client has an usual address private isn-t it ? or two eoip are needed for one public ip ?