Hi everybody.
An interesting question came across my desk - a customer of mine, is using
JunOS event policy (SRX to be exact), to capture change in his WAN IP
address and then populate it, in banch of other hierarchies.
The later logic worked as expected, but due to recent change, he switched
ISP and after we set up his new PPPoE dialer, we encountered the
following: the event-policy above, most of the time discovers pp0.0
interface IP address is either 0.0.0.0/0 or not-existing/empty.
After a short investigation, we identified the culprit: the event-policy
currently redponds to the snmp_trap_link_up event and sometimes, realy
sometimes, pp0.0 reaches up-up state, before an IP address was assigned.
"show interfaces terse" shows the interface state and it IP addresses,
exactly as above. So the event-policy it seems, does nothing wrong.
What's probably happening, is PPP Auth is complete (or some other state,
Juniper decided it enough to bring PPP interface into up-up state), while
IPCP isn't complete yet. Since the interface up, snmp_trap_link_up kicks in
and that's how we end up here.
Now, I don't think Juniper are in breach of something here (say, RFC). In
my humble opinion, I can see the logic in declaring a PPP interface up,
after say PPP Auth is complete, before IPCP (or another NCP, for that
matter) acquired address. Feel free correct me if I'm wrong. What I'm
"really" after: *is there an event corresponding to an interface acquiring
IP address?*
Cause, as I've said, I can see the distinction between interface state and
acquiring address (think of ethernet and DHCP) but than, how we're supposed
to capture that distinction in JunOS event policy?
Feel free to share you thoughts.
Thank you in advance,
Alex.
_______________________________________________
juniper-nsp mailing list juniper-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp
An interesting question came across my desk - a customer of mine, is using
JunOS event policy (SRX to be exact), to capture change in his WAN IP
address and then populate it, in banch of other hierarchies.
The later logic worked as expected, but due to recent change, he switched
ISP and after we set up his new PPPoE dialer, we encountered the
following: the event-policy above, most of the time discovers pp0.0
interface IP address is either 0.0.0.0/0 or not-existing/empty.
After a short investigation, we identified the culprit: the event-policy
currently redponds to the snmp_trap_link_up event and sometimes, realy
sometimes, pp0.0 reaches up-up state, before an IP address was assigned.
"show interfaces terse" shows the interface state and it IP addresses,
exactly as above. So the event-policy it seems, does nothing wrong.
What's probably happening, is PPP Auth is complete (or some other state,
Juniper decided it enough to bring PPP interface into up-up state), while
IPCP isn't complete yet. Since the interface up, snmp_trap_link_up kicks in
and that's how we end up here.
Now, I don't think Juniper are in breach of something here (say, RFC). In
my humble opinion, I can see the logic in declaring a PPP interface up,
after say PPP Auth is complete, before IPCP (or another NCP, for that
matter) acquired address. Feel free correct me if I'm wrong. What I'm
"really" after: *is there an event corresponding to an interface acquiring
IP address?*
Cause, as I've said, I can see the distinction between interface state and
acquiring address (think of ethernet and DHCP) but than, how we're supposed
to capture that distinction in JunOS event policy?
Feel free to share you thoughts.
Thank you in advance,
Alex.
_______________________________________________
juniper-nsp mailing list juniper-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp