Matus,

I have a question about the expected behavior for removing a
static mapping on an external interface address.

vpp# show int TenGigabitEthernet6/0/0 addr
TenGigabitEthernet6/0/0 (dn):
  L3 200.0.0.1/24
vpp# show nat44 static mappings
NAT44 static mappings:

And add a static mapping:

vpp# nat44 add static mapping  icmp local 10.10.10.100 external
TenGigabitEthernet6/0/0
vpp# show nat44 static mappings
NAT44 static mappings:
 local 10.10.10.100 external 200.0.0.1 vrf 0

Now, I tried to delete it:

vpp# nat44 add static mapping  icmp local 10.10.10.100 del
nat44 add static mapping: Mapping not exist.
vpp# nat44 add static mapping  icmp local 10.10.10.100 external
TenGigabitEthernet6/0/0 del
nat44 add static mapping: Mapping not exist.

I would have expected either of those two commands to have
removed the static mapping, but they do not.

This command, with an explicit external address does work:

vpp# nat44 add static mapping  icmp local 10.10.10.100 external 200.0.0.1
del
vpp# show nat44 static mappings
NAT44 static mappings:
vpp#

Is this the expected behavior here?  It seems to me that the mapping
should be able to be removed based on the (protocol, local-ip, local-port)
triplet and a little help from addr_only.

When it is not an interface address, that does seem to be sufficient:

vpp# nat44 add static mapping  icmp local 10.10.10.100
vpp# show nat44 static mapping
NAT44 static mappings:
 local 10.10.10.100 external 208.174.80.178 vrf 0
vpp# nat44 add static mapping  icmp local 10.10.10.100 del
vpp#

Furthermore, there apparently can not be multiple static mappings
for the same local-address:

vpp# nat44 add static mapping  icmp local 10.10.10.100 external
TenGigabitEthernet6/0/0
vpp# nat44 add static mapping  icmp local 10.10.10.100 external
TenGigabitEthernet6/0/0
nat44 add static mapping: Mapping already exist.
vpp#

So that triplet should make it a unique key for deletion, right?

Thanks,
jdl
  • [vpp-dev] Removin... Jon Loeliger
    • Re: [vpp-dev... Matus Fabian -X (matfabia - PANTHEON TECHNOLOGIES@Cisco)
      • Re: [vpp... Matus Fabian -X (matfabia - PANTHEON TECHNOLOGIES@Cisco)

Reply via email to