PIM sparse mode and auto-rp

Has anyone noticed any peculiarity with auto-rp and sparse mode. I 
expected that even though by default 224.0.1.39 and 224.0.1.40 are in 
dense mode, a mapping agent will only send the RP mappings to adjacent 
neighbors and that's why one could get away with auto-rp and sparse 
mode by having the MA and C-RP directly connected (or on the same 
router) and the MA connected to other PIM routers. 

With a topology as shown below, in which all the routers are setup for 
sparse mode, I was able to see the RP mappings on R2, R3 and R1 but 
not on R5 and R6. I understand typically I should use sparse-dense 
mode or autorp listener with sparse or even use autorp with sparse and 
static default RP but I just wanted to further understand how the RP- 
announce and discovery groups work in sparse mode by default. 

R3---serial-----R1-----Frame relay-------R5----------------R6 
                                             | 
                                             | 
                                            R2--------------------R4 

R4 was setup as C-RP and MA and I had the PIM nbma mode on R1's 
physical interface. R5 is using a point-to-point subinterface for 
connection to R1 and same goes for R2 over the frame relay network. 
All other links are ethernet. Is it something about sparse mode with 
autorp and serial point-to-point interfaces that is different over 
ethernet? 

Comments

  • If you are using only point-to-point frame-relay subinterfaces in R1, R2 and R5, PIM nbma mode shouldn't be needed.

    It seems like it should be working. My suggestion is follow the regular troubleshooting process.


    Check the mroute table, verify RPF accross the path, PIM state, etc

    If you are too tired because already tried everything and it doesn't seem to make sense anymore, try to reboot the machines before a new round of troubleshooting attempts.


    Cheers, Gustavo
Sign In or Register to comment.