[quagga-dev 7363] Re: Quagga 1.0?

Joakim Tjernlund joakim.tjernlund at transmode.se
Mon Nov 16 23:03:10 GMT 2009


Bartek Kania <mrbk at gnarf.org> wrote on 16/11/2009 21:01:17:
> >> Hi!
> >> Unfortunately it still crashes.
> >> And it logs the error you added. I've attached the log as
> >> log.txt. IPnumbers obscured tho.
> >> I've also attached my config and the full output from valgrind.
> > So some more debug:
>
> Hi!
> The "wrong nbrs" message appears very soon after start.
> The rn->lock message only appears after I shut down OSPFd, see partial
> logs below:
>
> For "wrong nbrs":
> 2009/11/16 20:54:55 OSPF: Packet[DD]: Neighbor x.y.B.255
> Negotiation done (Slave).
> 2009/11/16 20:54:55 OSPF: Packet[DD]: Neighbor x.y.B.245
> Negotiation done (Slave).
> 2009/11/16 20:54:55 OSPF: nsm_change_state(x.y.B.255, Loading ->
> Full): scheduling new router-LSA origination
> 2009/11/16 20:54:55 OSPF: nsm_change_state(x.y.B.245, Loading ->
> Full): scheduling new router-LSA origination
> 2009/11/16 20:54:57 OSPF: Wrong nbrs: 0x43bf550, 0x43a5dc0
> 2009/11/16 20:55:01 OSPF: Packet[DD]: Neighbor x.y.B.255
> Negotiation done (Slave).
> 2009/11/16 20:55:01 OSPF: nsm_change_state(x.y.B.255, Loading ->
> Full): scheduling new router-LSA origination
>
>
> After shutdown:
> 2009/11/16 20:58:02 OSPF: Terminating on signal
> 2009/11/16 20:58:02 OSPF: Wrong nbrs: 0x43c1b30, 0x43a5dc0
> 2009/11/16 20:58:02 OSPF: nsm_change_state(x.y.B.255, Full ->
> Deleted): scheduling new router-LSA origination
> 2009/11/16 20:58:02 OSPF: rn->lock! 3
> 2009/11/16 20:58:02 OSPF: Wrong nbrs: 0x441b9c8, 0x43a5dc0
> 2009/11/16 20:58:02 OSPF: nsm_change_state(x.y.B.255, Full ->
> Deleted): scheduling new router-LSA origination
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: Backup x.y.A.61
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: DR     x.y.A.61
> 2009/11/16 20:58:02 OSPF: rn->lock! 3
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: Backup x.y.A.61
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: DR     x.y.A.61
> 2009/11/16 20:58:02 OSPF: rn->lock! 3
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: Backup x.y.A.61
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: DR     x.y.A.61
> 2009/11/16 20:58:02 OSPF: rn->lock! 3
> 2009/11/16 20:58:02 OSPF: nsm_change_state(x.y.B.245, Full ->
> Deleted): scheduling new router-LSA origination
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: Backup x.y.A.56
> 2009/11/16 20:58:02 OSPF: DR-Election[1st]: DR     x.y.A.56
> 2009/11/16 20:58:02 OSPF: DR-Election[2nd]: Backup 0.0.0.0
> 2009/11/16 20:58:02 OSPF: DR-Election[2nd]: DR     x.y.A.56
> 2009/11/16 20:58:02 OSPF: interface x.y.A.56 [2] join AllDRouters
> Multicast group.
> 2009/11/16 20:58:02 OSPF: rn->lock! 3
> 2009/11/16 20:58:02 OSPF: interface x.y.A.56 [2] leave
> AllSPFRouters Multicast group.
> 2009/11/16 20:58:02 OSPF: interface x.y.A.56 [2] leave AllDRouters
> Multicast group.

I see that the second filed in the Wrong nbrs is always the same address,
wonder were that comes from? Perhaps you can improved the
logging here to see what nbrs are involved here?

Did you notice any change in behaviour? Does valgrind still complain?

the rn->lock! print out are not ideal but they happen when you shutdown
so it is not a big deal. I don't how they get there though.

 Jocke




More information about the Quagga-dev mailing list