From: Roberto Clavero Montano (Roberto.Clavero@TELINDUS.ES)
Date: Tue Jan 22 2008 - 09:51:03 ARST
Hi,
In real life scenario I have three Gig Ethernet rings, with almost 30 routers
in each ring speaking OSPF and PIM for redundancy. The three logical rings
have been connected to a central site that is routed by a Nokia nodes and
behind them the users watching real time camera.
Previous not working configs:
We have unicast traffic and multicast applications running and several users
watching TV properly. The drama came when the customer ask to us for a RP
redundancy.
We tried anycast RP redundancy using msdp but when it starts to sync then the
multicast traffic stops to arrive to the users. I had checked the logs and
there's nothing.
Doesn't make sense there's no message at Nokia or checkpoint tracker, so I
suppose that the nokias see the same IP via two MAC address and then them
start to drop all traffic. Along the rest of the ring the users can see join
to the groups even with anycast RP working, so the problem is at Nokia side.
So at this point, I had spoke with Nokia and they explain to me that dynamic
RPs will not work without downgrading the Catalyst.
MULTICASTUSERS
-----------R---------------------R-----------
| IGMP
-----------R---------------------R-----------
/ \
| /
\
R------< OSPF+PIM
------R---------NOKIAS+CHECKPOINTS--------R------< OSPF+PIM
------R
\ /
OSPF+PIM | OSPF+PIM \
/
-----------R---------------------R-----------
|
-----------R---------------------R-----------
| |
----
-----------------R----------------------- |
R
R---------------------------------
|
|
|
OSPF+PIM |
|
|
R
R
\----
-----------------------------------------/
|
R
My own workaround:
Because the problem came when the RP loopback with the same IP address in
other Catalyst is enabled, and at nokia's config is configured with static RP
address, I thought that I can provide the following workaround.
I will configure an IP SLA check for the prefix announced at the OSPF process
for the loopback who is the RP, and then when it stops to be announced by the
Catalyst because the fibber is down, or whatever then configures and action
thanks to eem to enable the loopback. Also I will configure other scripts to
shutdown the port when the primary restore the state (pre-empt emulation).
Did you find before some similar situation?
Thank you,
Roberto
Roberto Clavero Montano
CCNP, CCSP, SCSA, CCSA y WLFES
Seguretat de Xarxes, Telindus,S.A.U.
Parc de Negocis Mas Blau II
Avinguda de les Garrigues n: 38 - 44, planta baixa
Edifici Mar Blau
El Prat de Llobregat - 08820 Barcelona
Mailto:roberto.clavero@telindus.es
T +34 93 303 01 59
F +34 93 307 26 95
Telindus. Change things your way.
http://www.telindus.es
Internal Virus Database is out-of-date.
Checked by AVG Free Edition.
Version: 7.5.516 / Virus Database: 269.19.2/1223 - Release Date: 13/01/2008
20:23
This archive was generated by hypermail 2.1.4 : Fri Feb 01 2008 - 10:38:00 ARST