This bugs seems to still existing on 3.9-current :
Mar 12 17:46:20 core-vel-1 last message repeated 3 times
Mar 12 17:46:20 core-vel-1 bgpd[10577]: nexthop_update: non-existent nexthop
Mar 12 17:46:20 core-vel-1 bgpd[25242]: nexthop 85.31.106.1 now valid:
via 193.43.214.1
Mar 12 17:46:20 core-vel-1 bgpd[25242]: nexthop 85.31.106.1 now valid:
via 193.43.214.1
Mar 12 17:46:20 core-vel-1 bgpd[25242]: Lost child: route decision
engine terminated; signal 6
Mar 12 17:46:20 core-vel-1 bgpd[12319]: fatal in SE:
session_dispatch_imsg: pipe closed: Operation now in progress
Mar 12 17:46:26 core-vel-1 bgpd[25242]: kernel routing table decoupled
Mar 12 17:46:26 core-vel-1 bgpd[25242]: Terminating
OpenBSD core-vel-1.kazar.net 3.9 GENERIC.MP#598 i386
Please can this be investigated ? OpenBSD 3.8 doesn't have this nasty
bug....
/Xavier
Henning Brauer wrote:
that smells like a bad bug. I'll look into that asap.
* Xavier Beaudouin <[EMAIL PROTECTED]> [2006-02-14 11:26]:
Hi there,
I have in the two last snapshot (9/02 and 12/02) are exiting very
frequently with this error messages :
Feb 14 06:36:17 core-vel-1 bgpd[9573]: nexthop 85.xxx.xxx.1 now
valid: via 193.xx.xxx.1
Feb 14 06:36:17 core-vel-1 bgpd[20604]: fatal in RDE: nexthop_cmp:
unknown af
Feb 14 06:36:17 core-vel-1 bgpd[9573]: Lost child: route decision
engine exited
Feb 14 06:36:17 core-vel-1 bgpd[9831]: fatal in SE:
session_dispatch_imsg: pipe closed: Operation now in progress
Feb 14 06:36:19 core-vel-1 bgpd[9573]: kernel routing table decoupled
Feb 14 06:36:19 core-vel-1 bgpd[9573]: Terminating
This is really nasty because I lost the full mesh in *exaclty* same
time on two routers.
Previous snapshots didn't had this kind of behaviors...
Is there any way to add into bgpd a "sanity" to restart RDE when it
kill itself ?
/Xavier
