[j-nsp] NAPT pool "address" vs "address-range"
David B Funk
dbfunk at engineering.uiowa.edu
Mon Dec 28 23:28:45 EST 2015
When defining an address pool for NAPT you can use either
"address CIDR/blk" or "address-range low a.b.c.d high e.f.g.h"
and you'd think they'd do the same thing (with correct choice of values).
In an M120 w/ 10.4R8.5 they did but in an MX480 w/13.3R6.5 they don't.
Specifically with a config of:
[service nat]
pool 29-160-nat-pool {
address 128.255.29.160/28;
port {
automatic;
}
}
used in a NAT service
On the M120 you get a static route:
root at rtr-sc01b> show route protocol static
inet.0: 109 destinations, 117 routes (108 active, 0 holddown, 1 hidden)
+ = Active Route, - = Last Active, * = Both
[bunch of stuff omitted]
128.255.29.160/28 *[Static/1] 161w5d 03:26:09
Service to khaki-nat-service
But similar config ("address 128.255.20.160/28") on a MX480 get a static route:
root at rtr-sc02a> show route protocol static
inet.0: 57 destinations, 57 routes (56 active, 0 holddown, 1 hidden)
+ = Active Route, - = Last Active, * = Both
128.255.20.161/32 *[Static/1] 4d 02:03:58
Service to khaki-nat-service
128.255.20.162/31 *[Static/1] 4d 02:03:58
Service to khaki-nat-service
128.255.20.164/30 *[Static/1] 4d 02:03:58
Service to khaki-nat-service
128.255.20.168/30 *[Static/1] 4d 02:03:58
Service to khaki-nat-service
128.255.20.172/31 [Static/1] 00:00:14
Service to khaki-nat-service
128.255.20.174/32 [Static/1] 00:00:14
Service to khaki-nat-service
On the MX480 by changing that nat pool def to:
[service nat]
pool 29-160-nat-pool {
address-range low 128.255.20.160 high 128.255.20.175;
port {
automatic;
}
}
All of a sudden the static route becomes:
root at rtr-sc02a> show route protocol static
inet.0: 52 destinations, 52 routes (51 active, 0 holddown, 1 hidden)
+ = Active Route, - = Last Active, * = Both
128.255.20.160/28 *[Static/1] 00:05:24
Service to khaki-nat-service
Bottom line on the MX480 it's not willing to use the full range of the /28 CIDR block
for addresses in that pool, but by explicitly listing the full range it works.
The M120 didn't have this behaivor.
So what gives?
--
Dave Funk University of Iowa
<dbfunk (at) engineering.uiowa.edu> College of Engineering
319/335-5751 FAX: 319/384-0549 1256 Seamans Center
Sys_admin/Postmaster/cell_admin Iowa City, IA 52242-1527
#include <std_disclaimer.h>
Better is not better, 'standard' is better. B{
More information about the juniper-nsp
mailing list