RE: PIM NBMA Mode

From: Gupta, Gopal (NWCC) (gopal.gupta@hp.com)
Date: Sun Nov 11 2007 - 08:12:44 ART


That's where the issue is..my traffic is going to all
neighbors...completely lost into it.

Regards,
Gops

________________________________

From: George Goglidze [mailto:goglidze@gmail.com]
Sent: Sunday, November 11, 2007 16:40
To: Gupta, Gopal (NWCC)
Subject: Re: PIM NBMA Mode

Hi there,

Ignore my last post, I guess I was wrong.
here's what DocCD says:

Usage Guidelines

Use this command on Frame Relay, Switched Multimegabit Data Service
(SMDS), or ATM only, especially when these media do not have native
multicast available. Do not use this command on multicast-capable LANs
such as Ethernet or FDDI.

When this command is configured, each Protocol Independent Multicast
(PIM) join message is tracked in the outgoing interface list of a
multicast routing table entry. Therefore, only PIM WAN neighbors that
have joined for the group will get packets sent as data-link unicasts.
This command should only be used when the ip pim sparse-mode command is
configured on the interface. This command is not recommended for LANs
that have natural multicast capabilities.

So according to DocCD it should go unicast.

Regards,

On Nov 11, 2007 12:05 PM, George Goglidze <goglidze@gmail.com> wrote:

        Hi there,

        Actually as I understand it, pim nbma mode does not mean it will
no send traffic to all dlci-s.
        it just means that it keeps track of how many joins it receives,
and will cease sending traffic when it receives as many *part* messages
as it had *joins*.

        because on that kind of link, maybe it had before let's say 5
joins, and then one of the routers send part message, and it stops
sending traffic for all.
        so ip pim nbma mode was to prevent this.

        Maybe I'm wrong, check it in pim documentation,
        I dont have time to check it now.

        On Nov 11, 2007 11:37 AM, Gupta, Gopal (NWCC)
<gopal.gupta@hp.com > wrote:

                Hi Alex,

                As you know it is sparse mode it doesnt send traffic to
those who dont
                ask for means it is already pruned and then forward to
those who need
                the traffic.

                In this case R5 is thinking that join requests are
coming from serial
                interface so it is forwarding Traffic out to all DLCIs;
(Which NBMA mode
                command should have taken care for to send to only
155.1.0.3...that it
                has entry for in its multicast table.

                connection is like this

                R6-->SW1-->R3-->FRCloud<--R5

                 For R3

                (*, 232.2.2.2), 00:00:20/00:03:24, RP 150.1.4.4, flags:
SF
                 Incoming interface: Serial1/0.1, RPF nbr 155.1.0.5
                 Outgoing interface list:
                   Ethernet0/0, Forward/Sparse, 00:00:20/00:03:24

                ( 150.1.5.5 <http://150.1.5.5> , 232.2.2.2),
00:00:08/00:03:21, flags:
                 Incoming interface: Serial1/0.1, RPF nbr 155.1.0.5
                 Outgoing interface list:
                   Ethernet0/0, Forward/Sparse, 00:00:08/00:03:24

                SW1

                (*, 232.2.2.2), 01:28:34/stopped, RP 150.1.4.4, flags:
SJC
                 Incoming interface: Vlan3, RPF nbr 155.1.37.3
                 Outgoing interface list:
                   Vlan2, Forward/Sparse, 01:28:34/00:02:06

                ( 150.1.5.5 <http://150.1.5.5> , 232.2.2.2),
00:00:37/00:02:25, flags: JT
                 Incoming interface: Vlan3, RPF nbr 155.1.37.3
                 Outgoing interface list:
                   Vlan2, Forward/Sparse, 00:00:37/00:02:22

                (155.1.0.5, 232.2.2.2), 00:00:37/00:02:25, flags: JT
                 Incoming interface: Vlan3, RPF nbr 155.1.37.3
                 Outgoing interface list:
                   Vlan2, Forward/Sparse, 00:00:37/00:02:22

                R6 (Joined Group 232.2.2.2)

                *, 232.2.2.2), 01:30:58/stopped, RP 150.1.4.4, flags:
SJPL
                 Incoming interface: Ethernet0/0, RPF nbr 155.1.67.7
                 Outgoing interface list: Null

                (150.1.5.5, 232.2.2.2), 00:18:19/00:01:41, flags: PLT
                 Incoming interface: Ethernet0/0, RPF nbr 155.1.67.7
                 Outgoing interface list: Null

                (155.1.0.5, 232.2.2.2), 01:29:46/00:01:41, flags: PLT
                 Incoming interface: Ethernet0/0, RPF nbr 155.1.67.7
                 Outgoing interface list: Null

                (155.1.58.5, 232.2.2.2), 01:07:50/00:01:55, flags: PLT
                 Incoming interface: Ethernet0/0, RPF nbr 155.1.67.7
                 Outgoing interface list: Null

                Regards,

                Gops
                l Message-----
                From: Alex Steer [mailto:alex.steer@eison.co.uk]
                Sent: Sunday, November 11, 2007 15:52
                To: Gupta, Gopal (NWCC)
                Cc: Cisco certification
                Subject: RE: PIM NBMA Mode

                Hi Gops,

                It might be, it might not. Could you send a full copy
of your show ip
                mroute on all of the routers please (specifically the
S,G entry)?

                Nbma is primarily to stop a prune from one of the spokes
sent to r5 from
                pruning the other neighbors (default behaviour without
nbma mode). The
                other neighbors don't see the prune due to the
Non-broadcast nature of
                the interface and so don't prune override it. Pim nbma
causes pim to
                store network-hop information for that interface, that
doesn't mean it
                can unicast the packet to that 1 neighbor though, merely
means it wont
                prune the interface when it receives a prune from one of
the neighbors.
                It would be good for my own study to see the output.

                -----Original Message-----
                From: nobody@groupstudy.com [mailto:
nobody@groupstudy.com <mailto:nobody@groupstudy.com> ] On Behalf Of
                Gupta, Gopal (NWCC)
                Sent: 11 November 2007 09:56
                To: Cisco certification
                Subject: PIM NBMA Mode

                Hi Folks,

                There is issue with pim NBMA mode.

                R5 is the hub Router and sending traffic to group
232.2.2.2, The problem
                here is that despite of putting ip pim NBMA command on
R5 Serial
                interface, i am getting three replies; means R5 is still
sending traffic
                to all DLCIs configured under R5 serial interface.

                R5#ping 232.2.2.2

                Type escape sequence to abort.
                Sending 1, 100-byte ICMP Echos to 232.2.2.2, timeout is
2 seconds:

                Reply to request 0 from 155.1.67.6, 324 ms Reply to
request 0 from
                155.1.67.6, 728 ms Reply to request 0 from 155.1.67.6,
504 ms

                Debugs:-

                *Mar 1 11:26:16.277: Serial1/0(o): dlci 504(0x7C81),
pkt type
                0x800(IP), datagramsize 444 *Mar 1 11:26:16.277:
Serial1/0(o): dlci
                501(0x7C51), pkt type 0x800(IP), datagramsize 444 *Mar
1 11:26: 16.289:
                Serial1/0(o): dlci 503(0x7C71), pkt type 0x800(IP),
datagramsize 444

                These packets are packets for 232.2.2.2 address.

                R5#sh ip mr 232.2.2.2

                IP Multicast Routing Table

                (*, 232.2.2.2), 00:18:05/00:03:14, RP 150.1.5.5, flags:
S
                 Incoming interface: Null, RPF nbr 0.0.0.0
                 Outgoing interface list:
                   Serial1/0, 155.1.0.3, Forward/Sparse,
00:17:43/00:03:14

                Config :--

                interface Serial1/0
                 ip address 155.1.0.5 255.255.255.0
                 ip pim nbma-mode
                 ip pim sparse-mode
                 encapsulation frame-relay
                 no ip route-cache cef
                 no ip route-cache
                 serial restart-delay 0
                 no dce-terminal-timing-enable
                 frame-relay map ip 155.1.0.1 501 broadcast frame-relay
map ip
                155.1.0.3 503 broadcast frame-relay map ip 155.1.0.4
504 broadcast end

                Any Comments...how to overcome this, may i be i am
missing something
                here.

                Thanks
                Gops



This archive was generated by hypermail 2.1.4 : Sat Dec 01 2007 - 06:37:29 ART