Other device loses multicast when one device is disconnected

[PranavShekhar12762]

Hi, I have a cisco 3650 switch where 3 devices are connected to its tengig ports. All 3 ports are in same vlan and are recieving same multicats. PIM-SM is configured on this vlan. 

my problem is when I disconnect one device then other 2 devices loses the multicast and as soon as i reconnect the device rest other two reconnects the mutlicasts. What could be the issue? A quick advice will be helpfull. #3650 #PIM-SM #routing #multicast

[vb10]

1. So, is it collected, when multicast is broken, or when it's working? 

2. What IP addressing do you have in Vlan521? Interface seems to have different, address, than last reporters in this VLAN.  

Vlan521 is up, line protocol is up
Internet address is 10.0.52.254/24

239.1.201.3 Vlan521 17:44:31 00:02:41 10.0.51.15
239.1.201.5 Vlan521 00:10:53 00:02:37 10.0.51.11
239.1.201.7 Vlan521 17:44:31 00:02:40 10.0.51.15
239.1.201.6 Vlan521 00:10:53 00:02:32 10.0.51.11

In general, multicast stream might be stopped, if you remove reporter (host, which generated IGMP join for group), and other hosts don't report their interest in multicast for some reason. You need to check these outputs (show ip igmp groups and show ip igmp snooping groups) in broken state, to see, if there are active reporters. And if there no, need to check, whether all the decoders generate IGMP join messages.

[PranavShekhar12762]

Sorry for the confusion, this command was run when i changed the devices for testing. Actually its the same network.
10.0.52.15
10.0.52.11

[Jaderson Pessoa]

@PranavShekhar12762 , Hello

I think  that your RP has wrong configuration.

All multicasts shows the sources IP e.g.
(*, 239.1.204.7), 17:45:01/stopped, RP 1.1.1.4, flags: SJC
Incoming interface: GigabitEthernet1/0/7, RPF nbr 10.1.1.29
Outgoing interface list:
Vlan521, Forward/Sparse, 17:45:01/00:02:11

(10.0.14.117, 239.1.204.7), 17:45:01/00:01:24, flags: JT
Incoming interface: GigabitEthernet1/0/7, RPF nbr 10.1.1.29
Outgoing interface list:
Vlan521, Forward/Sparse, 17:45:01/00:02:11

I suggest to you configure correctly rendezvous point. If it is locally configure switch itself or other device that need be a RP or change to dense mode.

command to configure RP

ip pim rp-address

[PranavShekhar12762]

@Jaderson Pessoa I dont think the RPs are incorrect, because in that case device will not even receive the multicast. but here the case is different, the deviucs lose mutlicast when other devices in same vlan reboots or disconnected.
Others switches are also local.

[Jaderson Pessoa]

did you check who is 1.1.1.4?

[vb10]

This output looks correct. At this point it's not so important what RP address is, stream has already switched to SPT (shortest path tree), and might not even going via RP. Second entry (source, group) is being used now. 

If stream is working in general, and broken, whne one of the receivers are disconnected, then most likely problem is at the last segment, where receivers are located. 

(*, 239.1.204.7), 17:45:01/stopped, RP 1.1.1.4, flags: SJC
Incoming interface: GigabitEthernet1/0/7, RPF nbr 10.1.1.29
Outgoing interface list:
Vlan521, Forward/Sparse, 17:45:01/00:02:11

(10.0.14.117, 239.1.204.7), 17:45:01/00:01:24, flags: JT
Incoming interface: GigabitEthernet1/0/7, RPF nbr 10.1.1.29
Outgoing interface list:
Vlan521, Forward/Sparse, 17:45:01/00:02:11

[PranavShekhar12762]

OK, So any clue what could be wrong?

[PranavShekhar12762]

yes, that is RP in configured other switch from where 239.1.204.7 is coming with source IP 10.0.14.117 on this switch port gig1/0/7

[Georg Pauwen]

Hello,

if your IOS release supports it, check what happens when you configure 'ip igmp immediate-leave' on the SVI and/or the ports that receive the muticast traffic.

Also, which interfaces are the three hosts connected to ? I am trying to find out if they are all connected to the same, or different ASICs...

[PranavShekhar12762]

I have now found the root cause after several tests but don’t know whats the resolution,
As I said i have 3650 with 8 tengb ports, these 6 devices are connected to tengig1/1/1-6.
now every thing works fine when only 4 devices are connected. As soon as i connect 5th devices the problem starts.
does that mean cisco 3650 is not able to handle more than 4 devices on tengb ports? Can we resolve this with a config or its HW limitation?

[Georg Pauwen]

Hello,

you could try and change the IGMP version to 1 or 3 (2 is the default).

[PranavShekhar12762]

Do you mean just change IGMP version to 3 on every router?

[Georg Pauwen]

Hello,

indeed. Or version 1. I still think that the problem is related to the decoders themselves somehow. I guess you could test and verify that with connecting 'regular' multicast clients, if you have those available (PCs or laptops), and check if that same problem reoccurs.

[PranavShekhar12762]

Yes, no issues with 1GB ports when i connect other devices like PC. But problem seems to be in 10GB port. I changed the HWs as well the issue only start when more than 4 devices are connected to 10GB ports

[Georg Pauwen]

Hello,

are you using Cisco SFPs ? Post the output of 'sh ver'. Might be something simpler like a bug, or maybe you need a software upgrade to simultaneously support more than 4 10G SFPs...