GBLX router upgrade breaks bgp sessions

Wed Jul 10 14:56:12 UTC 2002

On Wed, Jul 10, 2002 at 07:04:38AM -0700, nanog wrote:
> Subject says it all.  GBLX upgraded some edge routers to a new JunOS
> release (possibly 5.3 rev 24)- and now our bgp sessions continually
> reset with:
> 
> Jul 10 06:58:24 MST: %BGP-3-NOTIFICATION: sent to neighbor X.X.X.X 3/3 (update missing required attributes) 0 bytes 
> 
> Anyone clueful at GBLX listening?  We've been down for about 4 hours,
> and the NOC (call center) people are less than helpful.

This sounds an awful lot like a problem we saw awhile back when
upgrading from JUNOS 4.x to 5.x.  At some point (I don't remember 
exactly when, but the details should be in the case notes of 
PR.19592) Juniper implemented a change that which makes their box
compliant with RFC 2858.  However, when speaking BGP with a 
non-RFC-compliant box (such as a Cisco running something like
12.0(15)S) the session flaps continuously in the manner you
describe because the other box expects the NEXT_HOP attribute to
be present in every update message.

Quoting from an email exchange I had with our Juniper rep:

"The result of the change is that JUNOS no longer sends the NEXT_HOP
attribute in an UPDATE message if only the MP_REACH_NLRI attribute is
present.  A workaround is to only use family inet unicast instead of
multicast or any on all BGP sessions to those cisco routers or upgrade
all of the cisco routers."

You might try forcing 'nlri uni' on your side to see if that 
does anything.  

Interested parties may wish to have a look at PR.22527, which was
opened at our request and adds a knob to revert back to the
non-RFC-compliant behavior, which is useful during a transition
period.

--Jeff