Thanks for the assistance everyone. This issue was resolved by shutting down a BGP peering session between Time Warner and Comcast. --Chad
Chad M. Reid, Network Administrator II
Work Hours: Sun. - Tue. 6AM-6PM and Wed. 6AM-3PM (MST -7)
Apollo Group | IT Services │ IT Operations Center (ITOC)
4025 S. Riverpoint Parkway │ MS: AA-M002 │ Phoenix, AZ 85040
phone: 602.557.6746 │ fax: 602.557.6606 │ email: chad.reid@apollogrp.edu
BGP Noob question here; but wouldn't Time Warner not recieve a prefix if it
wasn't reachable? Is this an artifact?
BGP Noob question here; but wouldn't Time Warner not recieve a prefix if it
wasn't reachable? Is this an artifact?
In a perfect world, people wouldn't advertise
prefixes unless they knew they had reachability
for those prefixes.
Unfortunately, this a far, far from perfect world.
Matt
It's a fair question and had nothing to do with the other network (TW
Telecom, in this case, not TIme Warner Cable).
Sorry for not filling in details sooner.
We recently needed to adjust the scale profile on some of our Cisco ASR9k
trident chip (80gig) Line Cards as we reached . The default profile only
supports up to 512k routes and will notify that CEF has run out of
DATA_TYPE_TABLE_SET space as a result. The profile change required (profile
13) requires a reload of the chassis. please be advised. You might want to
review with your local support team.
The symptom involved BGP neighbor instability and unreachability of some
destinations.
Shutting down that particular BGP session was a temporary work-around until
the adjustment could be made during a demand maintenance window to minimize
disruption.
Thanks,
Tony
Interesting; instability due to reaching this table limit? Good info!