Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Community Member

Multicast scalability

Hi,

I am running a network comprising of Catalyst 6513's with SUP7203B's. at present we have 800 VLAN's as we make use of a VLAN per access layer switch model.

I know have a problem that as soon as I enable multicast routing my SUP720's CPU runs at 100% and the system goes into a slowdown.

I would like to knwo where I can find information on the scalability of Multicast?

Regards

Dean

15 REPLIES
Hall of Fame Super Bronze

Re: Multicast scalability

Did you enable CEF ? IP Multicast Routing is done in hardware on the 6500s.

http://www.cisco.com/univercd/cc/td/doc/product/lan/cat6000/122sx/swcg/mcastv4.htm

Community Member

Re: Multicast scalability

Yes CEF is enabled. I get the following error in the log:

%MCAST-SP-4-MET_THRESHOLD_EXCEEDED: Multicast Expansion table has exceeded 98% of its capacity and is reaching its maximum

%MMLS-SP-6-MET_LIMIT_EXCEEDED: Failed to allocate MET entry, exceeded system limit of (65536) entries. Number of times MET limit is exceeded in the last 1 min : 430

Hall of Fame Super Bronze

Re: Multicast scalability

Did you enable PIM in all Vlans ? What mode did you select ? sparse or dense ?

Community Member

Re: Multicast scalability

Yes, I am using sparse mode.

Community Member

Re: Multicast scalability

Hi,

I assume that all access layer switches are layer 3 and are also running multicast -PIM. I also assume that all or a large number of the 800 VLANs are interested in the multicast traffic that you are sending. If both assumptions are true, then you will run into this issue because the core 6513 is replicating the traffic out a majority of those VLAN interfaces.

Basically a lot of interfaces in the OIL [out going interface list]. The best place to find information on multicast scalability on the 6500 would be to look at the prodcuts sheet of the Catalyst 6500 and the supervisor engine as well as the line casrds that you are using.

Try here:

http://www.cisco.com/en/US/products/hw/switches/ps708/products_data_sheets_list.html

Community Member

Re: Multicast scalability

Hi,

No all my access layer switches are Cisco 2950 / 2960 / 2970 L2 switches

Community Member

Re: Multicast scalability

Hi,

Having checked your error message in Cisco's Error message decoder:

##########################

1. %MMLS-6-MET_LIMIT_EXCEEDED: Failed to allocate MET entry, exceeded system limit of ([dec]) entries. Number of times MET limit is exceeded in the last 1 min : [dec]

This message indicates that the maximum MET entry limit has been exceeded. MET entries cannot be allocated by MMLS.

Recommended Action: The total number of OIFs is too large to fit in the MET table. Subsequent shortcuts or OIFs will be switched by the software. There is no workaround.

Related documents- No specific documents apply to this error message.

############################

2. %MMLS-6-MET_LIMIT_EXCEEDED: Failed to allocate MET entry, exceeded system limit of ([dec]) entries. Number of times MET limit is exceeded in the last 1 min : [dec]

This message indicates that the maximum MET entry limit has been exceeded. MET entries cannot be allocated by MMLS.

Recommended Action: The total number of OIFs is too large to fit in the MET table. Subsequent shortcuts or OIFs will be switched by the software. There is no workaround.

Related documents- No specific documents apply to this error message.

#######################

I suspect that there are routers hanging off the L2 switches that are running PIM and are interested in the multicast traffic. Can you confirm?

Community Member

Re: Multicast scalability

That is correct. Thye are running sparse mode.

Hall of Fame Super Bronze

Re: Multicast scalability

Did you configure a RP ?

Why do you need to enable PIM in all interfaces ?

Community Member

Re: Multicast scalability

Hi,

Yes I did. I presume it is best practise to do so?

sh ip pim rp

Group: 239.255.255.253, RP: 137.158.152.241, v2, next RP-reachable in 00:00:35

Group: 239.255.255.250, RP: 137.158.152.241, v2, next RP-reachable in 00:00:35

Group: 234.21.81.1, RP: 137.158.152.241, v2, next RP-reachable in 00:00:34

Group: 224.0.1.35, RP: 137.158.152.241, v2, next RP-reachable in 00:00:59

Group: 224.0.1.40, RP: 137.158.152.241, v2, next RP-reachable in 00:00:34

Group: 224.0.1.60, RP: 137.158.152.241, v2, next RP-reachable in 00:00:34

Group: 239.255.219.45, RP: 137.158.152.241, v2, next RP-reachable in 00:00:39

Group: 239.0.0.6, RP: 137.158.152.241, v2, next RP-reachable in 00:00:54

Group: 224.0.1.22, RP: 137.158.152.241, v2, next RP-reachable in 00:00:34

Community Member

Re: Multicast scalability

Hi,

We are using the Sup720-3B with around 200 Vlan configured for multicast. We have a MET table usage of 2%, so I expect for you a usage of less than 20%. I would enable multicast vlan by vlan and check when the MET increase unexpected. Than you could check with ip mroutes summary, where the OIF's are used. With show platform hardeware capacity you cloud check the MET.

router-720#show platform hardware capacity multicast

L3 Multicast Resources

IPv4 replication mode: ingress

IPv6 replication mode: ingress

Bi-directional PIM Designated Forwarder Table usage: 4 total, 0 (0%) used

Replication capability: Module IPv4 IPv6

1 ingress ingress

2 ingress ingress

3 ingress ingress

4 ingress ingress

5 ingress ingress

6 ingress ingress

7 egress egress

13 egress -

MET table Entries: Module Total Used %Used

7 65526 1060 2%

Community Member

Re: Multicast scalability

We removed multicast from a number of VLANS and below is the output after the change.

Router-720#show platform hardware capacity multicast

L3 Multicast Resources

IPv4 replication mode: egress

IPv6 replication mode: ingress

Bi-directional PIM Designated Forwarder Table usage: 4 total, 0 (0%) used

Replication capability:

Module IPv4 IPv6

7 egress egress

8 egress egress

9 egress egress

MET table Entries:

Module Total Used %Used

7 65526 1038 2%

8 65526 0 0%

#sh ip multicast

Multicast Routing: enabled

Multicast Multipath: disabled

Multicast Route limit: No limit

Multicast Triggered RPF check: enabled

Multicast Fallback group mode: Dense

Multicast DVMRP Interoperability: disabled

Community Member

Re: Multicast scalability

Hi,

A word of caution: One of the 'gotcha's' of multicast is that once you enable it natively, you have to enable it on every router in the network. If you don't do this you run the risk of breaking it. For e.g.

1) In the case of a failure when the RPF to a particular multicast source is routed via a different path [the one with no multicast enabled] on the even of a failure of a link or router

2) remember that on a LAN that conatins only routers [no receiving hosts- so no need for IGMP] if there is multicast feed on the LAN, all routers on that VLAN will receive a copy of the multicast feed because by default [without PIM snooping enabled on the switch] the switch will flood the multicast feed out all ports in that VLAN. This means that every router on the VLAN has to process switch this packet before they drop it [if they don't need it] this takes up CPU, whereas if multicast was enabled on all routers, the routers that don't need the feed will drop the packet less CPU utilization because their OIL for that group would ne null.

I guess what I'm trying to say is make sure that if you diable multicast on a VLAN, that VLAN will NEVER be used as the RPF to any multicast source in the event of any failure.

Community Member

Re: Multicast scalability

So I guess there are no error messages left. Now you have to enable multicast step by step on the vlan's, until the error messages apperars again. I guess there is a nasty client or something else connected to a vlan, which cause a lot of multicast entries.

Community Member

Re: Multicast scalability

Not to resurrect a dead thread, but we're having this same issue on several 6500's. We keep getting:

%MCAST-SP-4-MET_THRESHOLD_EXCEEDED: Multicast Expansion table has exceeded 98% of its capacity and is reaching its maximum

We have CEF enabled, PIM in sparce mode and an RP manually set.

(hostname omitted)#show platform hardware capacity multicast

L3 Multicast Resources

  IPv4 replication mode: egress

  IPv6 replication mode: egress

  Bi-directional PIM Designated Forwarder Table usage: 4 total, 0 (0%) used

  Replication capability: Module                              IPv4        IPv6

                          1                                 egress      egress

                          2                                 egress      egress

                          3                                 egress      egress

                          4                                 egress      egress

                          5                                 egress      egress

                          6                                 egress      egress

                          7                                 egress      egress

                          8                                 egress      egress

  MET table Entries: Module                             Total    Used    %Used

                     5                                  65516   65340     100%

                     6                                  65516       0       0%

Multicast LTL Resources

  Usage:   30656 Total, 1174 Used

I have read several threads and have not seen a solution to this. I guess the only difference we have from the original post is that CPU usage is not high. We had CEF off at one point and the IP Input was causing CPU to be high (understanably), but upon turning CEF back on it normalized. Thoughts?

1643
Views
0
Helpful
15
Replies
CreatePlease to create content