Cisco IOS multicast bug and workarounds

Bill Nickless nickless at mcs.anl.gov
Fri Feb 25 18:32:11 CST 2000


If your local network people use Cisco routers to supply IP multicast 
service to your Access Grid node, or if you are using a Cisco 2621 router 
to "roll your own" solution, please read this message and pass it along as 
appropriate.

There is a bug in Cisco IOS that can cause your source not to be advertised 
properly outside your site.  If you're registered with Cisco for support, 
you can look at this page for all the details:

   http://www.cisco.com/cgi-bin/Support/Bugtool/onebug.pl?bugid=CSCdp68820

Here is a summary of the problem:

     Sources of multicast packets that send at a low frequency
     may have their (S,G) expired in the first-hop designated router (DR).
     The corresponding (S,G) may also expire in the RP.
     If the (S,G) state is later created from a PIM Join packet from
     down stream, then the register flag is not set for this (S,G)
     in this router.  If a packet is then received from the source,
     register packet would not be sent to the Rendevouz Point (RP).
     Therefore, the A-flag for the (S,G) will not be set in the RP
     which would prevent the subject source from being advertised
     to the MSDP peers.

The fix for this is available in Cisco IOS 12.0(9)S which can be downloaded 
and run today on medium and high-end routers (unfortunately excluding the 
2621).  This version, of course, can only be run when a site only uses IP 
on the router involved.

If you need other protocols, or if you want this fix on your 2621 router, 
you'll have to wait for 12.0(10) which is scheduled for release on 
27-Mar-00.  I do not yet know what the memory requirements will be for the 
12.0(10) release on the 2621 router, but will follow up as soon as I learn 
if we need to upgrade the 2621 routers we have deployed for "roll your own" 
multicast situations.

You're being affected by this problem if you are transmitting but nobody 
outside your organization sees you.  This is especially true if people are 
seeing your transmission from other machines on your network.

To confirm that you're seeing the problem, type the following at your Cisco 
router directly connected to your host that is transmitting:

    'show ip mroute <group> <host>'

You should see something like this:

   destiny#show ip mroute 224.2.177.156 140.221.9.223 sum
   IP Multicast Routing Table
   Flags: D - Dense, S - Sparse, C - Connected, L - Local, P - Pruned
          R - RP-bit set, F - Register flag, T - SPT-bit set, J - Join SPT
          M - MSDP created entry, X - Proxy Join Timer Running
          A - Advertised via MSDP
   Timers: Uptime/Expires
   Interface state: Interface, Next-Hop or VCD, State/Mode

     (140.221.9.223, 224.2.177.156), 03:01:55/00:03:29, flags: CFT
                                                                ^
                                                                |
                      **** IMPORTANT FLAG TO LOOK FOR **** ------

If you do NOT see the F flag, then you are suffering from the problem!

To fix: type 'clear ip mroute <group number> <host>' on the Cisco router 
that your transmitting host is directly attached to.  In a minute or two 
you will see the F flag appear, and your traffic should start flowing out 
to the rest of the Access Grid sites.

As always, please feel free to contact me if you have any questions or 
concerns about this issue.  I will be happy to work with you in whatever 
way I can to ensure your IP multicast environment is working cleanly.
===
Bill Nickless    http://www.mcs.anl.gov/people/nickless      +1 630 252 7390
PGP:0E 0F 16 80 C5 B1 69 52 E1 44 1A A5 0E 1B 74 F7     nickless at mcs.anl.gov




More information about the ag-tech mailing list